consumer confusion: a test of the behavioural …orca.cf.ac.uk/56902/1/2013 anninou i.pdf ·...

CONSUMER CONFUSION: A TEST OF THE

BEHAVIOURAL PERSPECTIVE MODEL

by

Ioanna Anninou

A Thesis Submitted in Fulfilment of the Requirements for the Degree of Doctor of

Philosophy of Cardiff University

Marketing and Strategy Section of Cardiff Business School,

Cardiff University

September 2013

ii

DECLARATION

This work has not previously been accepted in substance for any degree and

is not concurrently submitted in candidature for any degree.

Signed ……… ……………………. (Ioanna Anninou)

Date ……16 September 2013…..

STATEMENT 1

This thesis is being submitted in partial fulfilment of the requirements for the degree of

PhD.

Signed ……… ……………….…. (Ioanna Anninou)

Date ……16 September 2013…...

STATEMENT 2

This thesis is the result of my own independent work/investigation, except where

otherwise stated.

Other sources are acknowledged by footnotes giving explicit references.

Signed ……… …………………. (Ioanna Anninou)


STATEMENT 3

I hereby give consent for my thesis, if accepted, to be available for photocopying and for

inter-library loan, and for the title and summary to be made available to outside

organisations.

Signed ……… ………………… (Ioanna Anninou)


iii

ACKNOWLEDGEMENTS

This study has been the result of both personal and academic interest towards consumer

emotions (especially the case of consumer confusion) and rule-governed behaviour. It is

the product of discussions (on practical and theoretical aspects) with both Professor

Gordon Foxall and Dr. John Pallister, who as my supervisors, I would like to deeply thank

for their support throughout this project. They have both been great advocates of this

attempt, offered constructive criticism, corrections and encouragement, in order for the

final form of this thesis to be realised.

It goes without saying that the greatest gratitude for the implementation of this task should

be attributed to the research participants, who contributed to the quantitative study and

pilot tests and spent some of their personal time, views and in many cases valuable

resources for the completion of this task. I would also like to thank both the account and

project manager of the online research panel for the great help they offered with this

project.

Dr. Mirella Yani-de-Soriano supported this study especially with comments on the first

drafts of the questionnaire and the quantitative analysis of the data. Dr. I. Christodoulou

and Dr. A. Piterou have also facilitated the process of pilot test data collection and I would

like to express my gratitude to them. Several academics who participated in the 2nd

Biennial AMS Doctoral Consortium commended on early proposals of the thesis and it

has been a pleasure meeting them.

Both Lainey Clayton and Elsie Philips from the PhD office have always been kind and

helpful, even when times were difficult. My warm thanks also go to other PhD students

(friends) at Cardiff Business School and especially my flatmate at Cardiff, Anastasia, with

whom we have taken this long PhD journey together.

Last and foremost a great thank you to my family. Especially my parents, my sister, my

brother-in-law and nephew have given me their unequivocal support. My husband, Theo,

had to deal with all of my daily problems and times that my confidence was tested.

Without his support this study would not be realised. No words seem enough to express

my gratitude to him.

iv

ABSTRACT

In light of the increasing academic and practical importance of consumer confusion,

more theoretical and empirical inquiries are necessary in order to comprehend this

concept. This study extends the notion of confusion by adopting the idea of self-

based rules. Confusion can be defined as a self-based track (or better a rule for the

lack of rules and norms, a special case of anomy). As a rule, there is a differing

language that can be used to describe it– the first, extensional, deals with confusion

as an overall response to physical and social stimuli and the other, intentional, deals

with it in terms of individual understanding and beliefs. This study uses the

theoretical principles of the Behavioural Perspective Model (BPM) as its primary

device. The current state of the BPM dictates the use of an extensional language

(BPM-E). The model will be extended and placed within the framework and study

of an intentional explanation (BPM-I). The explanatory or interpretative role that

confusion can play in these models will be described. Specific research hypotheses

that correspond to these explanations have been developed.

In order to implement these objectives a main quantitative survey (N=260) which

provided data on 520 consumer situations, has been informed by a meaningful in the

produced results pilot-exploratory study (N=7) and multiple qualitative (N=10) and

quantitative (N=56) pilot tests and discussions with knowledgeable and lay

participants. Multiple regression and ANOVA indicate significant main effects

when Mehrabian and Russell’s affective scales and different kinds of confusion

(similarity-complexity) are used to predict approach-avoidance behavioural

responses. Additionally, support is provided for the patterns expected from the

affective and behavioural variables when these are applied to other situations

beyond the original eight contingency categories of the BPM.

The main contribution of this study lies with the inclusion of an aversive

consequence of shopping situations in the BPM and the extension of the model

towards embracing and applying intentionality. Overall, this study supports the

supposition put forward by Foxall (2004; 2007a; 2007b; 2013) that the intentional

BPM can add and extend the explanatory power of the extensional model.

KEYWORDS: Retail settings, Behavioural Perspective Model (BPM), intentional

behaviourism, utilitarian and informational reinforcement, consumer confusion,

rule-governed behaviour, self-based rules.

v

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ........................................................................................................................ III

ABSTRACT ..................................................................................................................................................IV

TABLE OF CONTENTS .............................................................................................................................. V

LIST OF TABLES ........................................................................................................................................ X

LIST OF FIGURES ................................................................................................................................. XIII

ABBREVIATIONS ................................................................................................................................... XIV

1. INTRODUCTION ................................................................................................................................ 1

1.1. Introduction ..................................................................................................................................... 1

1.2. Research Background ..................................................................................................................... 1

1.2.1. Sustainable Retailing– Environmental Concerns ......................................................................... 2

1.2.2. Multichannel Shopping on the Rise ............................................................................................. 3

1.2.3. Changing Consumer Behaviour ................................................................................................... 4

1.2.4. Retailtainment .............................................................................................................................. 5

1.2.5. Market Dominance ...................................................................................................................... 5

1.2.6. Overall Evaluation of the Retail Environment ............................................................................. 6

1.3. Research Problem ........................................................................................................................... 9

1.4. Research Purpose and Objectives ............................................................................................... 13

1.5. Research Hypotheses .................................................................................................................... 14

1.6. Justification of this Research ....................................................................................................... 16

1.7. Research Methodology ................................................................................................................. 17

1.8. The Contributions of this Thesis.................................................................................................. 18

1.9. Thesis Structure and Order ......................................................................................................... 19

1.10. Conclusion ..................................................................................................................................... 22

2. THE NATURE OF CONFUSION IN PSYCHOLOGICAL RESEARCH ................................... 23

2.1. Introduction ................................................................................................................................... 23

2.2. Previous Treatments of Emotional Words ................................................................................. 23

2.3. Separating Emotions from Other Entities .................................................................................. 26

2.3.1. The Classical View of a Concept ............................................................................................... 27

2.3.2. Everyday Use of Language ........................................................................................................ 28

2.3.3. Criteria for Measuring Emotions ............................................................................................... 28

2.4. The State of Confusion ................................................................................................................. 30

2.4.1. Confusion as a Cognitive State .................................................................................................. 30

2.4.2. Confusion as a Cognitive Feeling .............................................................................................. 31

2.4.3. Confusion as an Emotion ........................................................................................................... 32

2.5. Re-examining the Debate on Confusion ...................................................................................... 32

2.6. Conclusion ..................................................................................................................................... 35

vi

3. THE TREATMENT OF CONFUSION IN CONSUMER BEHAVIOUR LITERATURE ......... 36

3.1. Introduction ................................................................................................................................... 36

3.2. Seeking Variety and ‘Extraordinary Experiences’ .................................................................... 36

3.3. Some Opposing Arguments .......................................................................................................... 39

3.4. Consumer Confusion .................................................................................................................... 41

3.4.1. Initial Conceptions of Confusion ............................................................................................... 42

3.4.2. Deeper Investigations of Confusion .......................................................................................... 47

3.5. Consequences of Consumer Confusion ....................................................................................... 55

3.6. Demarcation of the Terms ............................................................................................................ 56

3.6.1. Cognitive Dissonance ................................................................................................................ 56

3.6.2. Buying Risk and Uncertainty ..................................................................................................... 57

3.6.3. Irritation/Frustration .................................................................................................................. 59

3.7. Conclusion ..................................................................................................................................... 59

4. THE BPM AND INTENTIONAL BEHAVIOURISM ................................................................... 61

4.1. Introduction ................................................................................................................................... 61

4.2. The Explanation of Consumer Behaviour .................................................................................. 61

4.3. Ways of Learning (Cognitive versus Behavioural) .................................................................... 64

4.4. Behaviour Analysis ....................................................................................................................... 66

4.5. Operant Conditioning ................................................................................................................... 69

4.6. Kinds of Reinforcement (and Punishment) ................................................................................ 71

4.6.1. Primary and Secondary Reinforcers .......................................................................................... 71

4.6.2. Contingency-Derived and Rule-Derived Reinforcers ................................................................ 71

4.6.3. Utilitarian and Informational Reinforcers .................................................................................. 78

4.7. The Contextual Stance .................................................................................................................. 79

4.8. The BPM (The Behavioural Perspective Model) ........................................................................ 80

4.8.1. Level I: The Operant Class ........................................................................................................ 82

4.8.2. Level II: The Contingency Category ......................................................................................... 83

4.8.3. Level III: The Consumer Situation ............................................................................................ 86

4.9. Cognitive Psychology .................................................................................................................... 87

4.10. The Intentional Stance .................................................................................................................. 89

4.11. Dennett’s Intentionality ................................................................................................................ 91

4.12. Intentional Behaviourism ............................................................................................................. 93

4.13. Conclusion ..................................................................................................................................... 96

5. MEHRABIAN AND RUSSELL’S ‘APPROACH TO ENVIRONMENTAL PSYCHOLOGY’ . 97

5.1. Introduction ................................................................................................................................... 97

5.2. Categories of Emotional Theories and Consumer Behaviour ................................................... 97

5.3. Mehrabian and Russell’s Approach .......................................................................................... 101

vii

5.4. The Emotional Dimensions ........................................................................................................ 103

5.5. Approach- Avoidance Behaviour .............................................................................................. 106

5.6. Applying the Mehrabian and Russell Approach ...................................................................... 108

5.7. A Role for Dominance ................................................................................................................ 112

5.8. Conclusion ................................................................................................................................... 114

6. CONCEPTUAL FRAMEWORK ................................................................................................... 116

6.1. Introduction ................................................................................................................................. 116

6.2. The State of Confusion ............................................................................................................... 117

6.3. Confusion as a Self-Based Rule.................................................................................................. 118

6.4. Confusion as ‘Anomy’ ................................................................................................................ 119

6.5. Confusion as a Self-Based Rule (or ‘A Rule for the Lack of Rules’) ...................................... 120

6.6. The Languages of Explanation .................................................................................................. 121

6.7. The Behavioural Perspective Model Extensional– (BPM-E) .................................................. 124

6.8. The Behavioural Perspective Model Intentional– (BPM-I) ..................................................... 129

6.8.1. The Debate on the Cognitive-Affective (or Binary) Nature of Confusion............................... 131

6.8.2. The Conscious-Subconscious Debate ...................................................................................... 131

6.9. Main Proposition of this Thesis ................................................................................................. 133

6.10. Development of Research Hypotheses ....................................................................................... 135

6.10.1. Group of Hypotheses 1 ............................................................................................................ 135


6.10.3. Choice of Situations Examined................................................................................................ 142



6.11. Conclusion ................................................................................................................................... 149

7. RESEARCH METHODOLOGY ................................................................................................... 151

7.1. Introduction ................................................................................................................................. 151

7.2. Scientific Research Paradigm .................................................................................................... 151

7.2.1. Overview of Research Paradigms and Guiding Beliefs of this Inquiry ................................... 152

7.2.2. Philosophical Position ............................................................................................................. 156

7.3. Research Design .......................................................................................................................... 158

7.3.1. Research Process ..................................................................................................................... 158

7.3.2. Research Purpose ..................................................................................................................... 159

7.3.3. Data Collection Timeframe ..................................................................................................... 165

7.4. Research Methods ....................................................................................................................... 166

7.4.1. Research Phase 1- Critical Evaluation of the Literature .......................................................... 168

7.4.2. Research Phase 2 and 3- Use of Different Kinds of Interviews ............................................... 168

7.4.3. Interviews ................................................................................................................................ 170

7.4.4. Exploratory (Pilot) Test ........................................................................................................... 173

viii

7.4.5. Pilot Tests ................................................................................................................................ 180

7.4.6. The Research Instrument- Questionnaire ................................................................................. 184

7.4.7. Sample Procedure and Size...................................................................................................... 188

7.4.8. Data Collection Procedures ..................................................................................................... 190

7.4.9. Praxiology ................................................................................................................................ 194

7.5. Data Analysis Techniques .......................................................................................................... 198

7.5.1. Hypothesis Testing .................................................................................................................. 198

7.5.2. Descriptive and Uni-Bi-Multi-Variate Techniques .................................................................. 199

7.6. Research Integrity and Quality Criteria ................................................................................... 201

7.6.1. Quality Criteria for Judging the Research as a Whole ............................................................. 202

7.6.2. Specific Quality Criteria for Judging the Quantitative Constructs .......................................... 205

7.7. Data Preparation for Analysis ................................................................................................... 207

7.8. Conclusion ................................................................................................................................... 208

8. DATA ANALYSIS ........................................................................................................................... 210

8.1. Introduction ................................................................................................................................. 210

8.2. Sample Demographic Profile ..................................................................................................... 211

8.3. Preliminary Analysis .................................................................................................................. 214

8.3.1. Missing Data ............................................................................................................................ 214

8.3.2. Outliers (Univariate and Multivariate) ..................................................................................... 217

8.3.3. Normality (Univariate and Multivariate) ................................................................................. 219

8.4. Descriptive Analysis .................................................................................................................... 222

8.4.1. Means (Sd)/ Skewness/ Kurtosis of Individual Items .............................................................. 222

8.5. Psychometric Properties of the Scales ....................................................................................... 224

8.5.1. Dimensionality ......................................................................................................................... 225

8.5.2. Reliability ................................................................................................................................ 232

8.6. Hypothesis Testing ...................................................................................................................... 234

8.6.1. Examining the BPM-E (The Extensional Behavioural Perspective Model) ............................ 234

8.6.2. Examining the BPM-I (The Intentional Behavioural Perspective Model) ............................... 237

8.6.3. The Relationships Between the Variables ............................................................................... 239

8.6.4. The Effect of Consumer Confusion ......................................................................................... 244

8.6.5. The Degree of the Relationship between the Affective-Confusion and Behavioural Variables

249

8.6.6. The Pattern of the Relationship between the Affective/Confusion Variables and Aminusa.... 261

8.7. The Results of the Statistical Hypothesis Testing ..................................................................... 263

8.8. Conclusion ................................................................................................................................... 267

9. FURTHER ANALYSIS ................................................................................................................... 268

9.1. Introduction ................................................................................................................................. 268

9.2. Intentional Confusion and Consumer Socio-Demographic Characteristics .......................... 268

ix

9.3. The Scale of Confusion ............................................................................................................... 273

9.4. The PAD and Behavioural Variables ........................................................................................ 274

9.5. Placing This Study’s Situations in the Context of the BPM Operant Classes and Contingency

Categories .................................................................................................................................................... 276

9.6. Conclusion ................................................................................................................................... 283

10. DISCUSSION, IMPLICATIONS AND FUTURE RESEARCH ............................................ 284

10.1. Introduction ................................................................................................................................. 284

10.2. Research Overview ..................................................................................................................... 284

10.3. Theoretical Findings and Implications ..................................................................................... 288

10.3.1. The Nature of Confusion ......................................................................................................... 288

10.3.2. The Conceptualisation of Confusion ....................................................................................... 290

10.3.3. The Contextual Treatment of Confusion ................................................................................. 293

10.3.4. The Extensional Model (BPM-E) ............................................................................................ 295

10.3.5. The Intentional Model (BPM-I) ............................................................................................... 298

The BPM and the MR Approach .............................................................................................................. 303

10.3.6. The Role of Dominance ........................................................................................................... 304

10.4. Managerial Implications ............................................................................................................ 310

10.5. Contributions of this Thesis ....................................................................................................... 314

10.5.1. Contributions by Chapter ......................................................................................................... 314

10.5.2. Objective Driven (and Overall) Contribution .......................................................................... 317

10.6. Limitations and Directions for Future Research ..................................................................... 318

10.6.1. Limitations ............................................................................................................................... 320

10.6.2. Directions for Future Research ................................................................................................ 322

10.7. Conclusion and Reflections on the Study .................................................................................. 324

11. REFERENCES............................................................................................................................ 328

12. APPENDIX .................................................................................................................................. 361

x

LIST OF TABLES

TABLE 1.1 RESEARCH HYPOTHESES OF THIS STUDY ....................................................................... 15

TABLE 2.1 173 NEGATIVE AND 143 POSITIVE EMOTION WORDS ................................................... 24

TABLE 3.1THE 'ENVIRONMENTAL APPROACH' TO CONSUMER CONFUSION ............................. 52

TABLE 3.2 THE ‘COGNITIVE APPROACH' TO CONSUMER CONFUSION ........................................ 54

TABLE 4.1 FACTORS AFFECTING CONSUMER BEHAVIOUR ............................................................ 62

TABLE 4.2 BEHAVIOURAL AND COGNITIVE APPROACHES TO DECISION-MAKING ................. 77

TABLE 4.3 SOURCES OF REINFORCEMENT .......................................................................................... 79

TABLE 4.4 OPERANT CLASSIFICATION OF CONSUMER BEHAVIOUR ........................................... 82

TABLE 4.5 THE BPM CONTINGENCY MATRIX..................................................................................... 85

TABLE 5.1 OVERVIEW OF PUBLISHED CONSUMER BEHAVIOUR STUDIES USING EMOTIONS

AS MAIN VARIABLES ..................................................................................................................... 100

TABLE 5.2 THE SEMANTIC DIFFERENTIAL MEASURES OF EMOTIONAL STATES .................... 105

TABLE 5.3 THE FOUR DIMENSIONS OF APPROACH-AVOIDANCE BEHAVIOUR ........................ 107

TABLE 6.1 A DEPICTION OF THE MAIN CHARACTERISTICS (‘QUALIA’) OF THE STATE OF

CONFUSION ...................................................................................................................................... 117

TABLE 6.2 THE NATURE OF CONFUSION AT THE TWO LEVELS (EXTENSIONAL AND

INTENTIONAL) PROPOSED BY THIS STUDY ............................................................................. 134

TABLE 6.3 BEHAVIOURAL AND COGNITIVE APPROACHES TO DECISION-MAKING. .............. 141

TABLE 7.1 FOUR SOCIAL RESEARCH PARADIGMS EXPLAINED ................................................... 153

TABLE 7.2 ONTOLOGY, EPISTEMOLOGY AND RESEARCH METHODS OF THIS STUDY .......... 157

TABLE 7.3 SUMMARY OF RESEARCH OBJECTIVES AND QUESTIONS OF EACH RESEARCH

STAGE ................................................................................................................................................ 162

TABLE 7.4 DETAILS OF RESEARCH METHODS USED IN THIS STUDY FOR THE PURPOSES OF

QUESTIONNAIRE DEVELOPMENT, TESTING AND DATA COLLECTION. ............................ 170

TABLE 7.5 PARTICIPANTS OF THE EXPLORATORY TEST ............................................................... 174

TABLE 7.6 SPONTANEOUS ANSWERS ON THE TOPIC OF 'CONSUMER CONFUSION'

ACCORDING TO THE PARTICIPANTS ......................................................................................... 176

TABLE 7.7 ADVANTAGES AND DISADVANTAGES OF ONLINE RESEARCH PANELS FOR

(ACADEMIC) RESEARCH ............................................................................................................... 192

TABLE 7.8 REVERSED ITEMS OF THE PAD SCALE ........................................................................... 207

TABLE 8.1 SAMPLE DEMOGRAPHIC PROFILE ................................................................................... 213

TABLE 8.2 DESCRIPTIVE STATISTICS OF INDIVIDUAL ITEMS ...................................................... 223

TABLE 8.3 KMO AND BTS MEASURE OF SAMPLING ADEQUACY. ALL ITEMS .......................... 228

TABLE 8.4 TOTAL VARIANCE EXPLAINED ........................................................................................ 228

TABLE 8.5 KMO AND BTS MEASURE OF SAMPLING ADEQUACY. EMOTIONAL VARIABLES 229

xi

TABLE 8.6 FACTOR ANALYSIS (VARIMAX ROTATION) OF THE AFFECTIVE AND

BEHAVIOURAL MEASURES .......................................................................................................... 229

TABLE 8.7 KMO AND BTS OF SAMPLING ADEQUACY. CONFUSION ........................................... 231

TABLE 8.8 FACTOR ANALYSIS (OBLIQUE ROTATION) OF THE CONFUSION MEASURES ....... 231

TABLE 8.9 RELIABILITY (CRONBACH'S ALPHA) FOR THE SCALES OF THIS STUDY ............... 233

TABLE 8.10 DESCRIPTIVE STATISTICS OF COMPOSITE VARIABLES ........................................... 234

TABLE 8.11 MEANS AND ANOVA RESULTS FOR THE TWO SITUATIONS (N=260 FOR EACH

MARKET) ........................................................................................................................................... 235

TABLE 8.12 MEAN AND MEDIAN FOR THE TWO CONFUSION VARIABLES ............................... 237

TABLE 8.13 MEANS AND ANOVA RESULTS FOR SIMILARITY CONFUSION (BPM-I) ................ 237

TABLE 8.14 MEANS AND ANOVA RESULTS FOR COMPLEXITY CONFUSION (BPM-I) ............. 238

TABLE 8.15 INTERPRETATION OF PEARSON CORRELATION (R) .................................................. 241

TABLE 8.16 CORRELATION COEFFICIENTS OF THIS STUDY (N=520) ........................................... 242

TABLE 8.17 CORRELATION COEFFICIENTS BETWEEN KINDS OF CONFUSION IN THE TWO

MARKETS .......................................................................................................................................... 243

TABLE 8.18 CORRELATION COEFFICIENTS PER SITUATION ......................................................... 244

TABLE 8.19 FACTORIAL ANOVA RESULTS FOR SIMILARITY CONFUSION ................................ 246

TABLE 8.20 FACTORIAL ANOVA RESULTS FOR COMPLEXITY CONFUSION ............................. 247

TABLE 8.21CELL MEANS FOR PLEASURE (COMPLEXITY-SITUATIONS) .................................... 248

TABLE 8.22 CELL MEANS FOR DOMINANCE (COMPLEXITY- SITUATIONS) .............................. 248

TABLE 8.23CELL MEANS FOR APPROACH (COMPLEXITY-SITUATIONS) ................................... 249

TABLE 8.24 REGRESSION FOR APPROACH BEHAVIOUR, N=520 ................................................... 258

TABLE 8.25 REGRESSION FOR AVOIDANCE BEHAVIOUR, N=520 ................................................. 259

TABLE 8.26 REGRESSION FOR AMINUSA BEHAVIOUR, N=520 ...................................................... 260

TABLE 8.27 ANOVA RESULTS FOR APPROACH-AVOIDANCE (AMINUSA) PAD+SIMILARITY.

............................................................................................................................................................. 262

TABLE 8.28 PLEASURE AND AROUSAL INTERACTION. CELL MEANS FOR AMINUSA ............ 262

TABLE 8.29 ANOVA RESULTS FOR APPROACH-AVOIDANCE (AMINUSA) PAD+COMPLEXITY

............................................................................................................................................................. 263

TABLE 8.30 RESULTS OF STATISTICAL HYPOTHESIS TESTING .................................................... 265

TABLE 9.1 ANOVA RESULTS FOR SOCIO-DEMOGRAPHIC CHARACTERISTICS AND LEVELS

OF CONFUSION IN THE TWO MARKETS .................................................................................... 271

TABLE 9.2 A COMPARISON OF THE INTERNAL CONSISTENCY OF THE CONFUSION

MEASURES ........................................................................................................................................ 273

TABLE 9.3 A COMPARISON OF THE CORRELATIONS OF THE CONFUSION MEASURES ......... 273

TABLE 9.4 A COMPARISON OF THE INTERNAL CONSISTENCY OF THE MEASURES ............... 275

TABLE 9.5 A COMPARISON OF THE CORRELATIONS OF THE PAD MEASURES ........................ 275

TABLE 9.6 CONCEPTUAL COMPARISON OF THE TWO RETAIL SITUATIONS (BASED ON BPM-

E) ......................................................................................................................................................... 277

xii

TABLE 9.7 RANGE AND MEANS OF PAD AND AMINUSA FOR THE EIGHT PUBLISHED STUDIES

OF THE BPM ...................................................................................................................................... 279

TABLE 9.8 MEANS FOR THE TWO SITUATIONS EXAMINED IN THIS STUDY. GROCERY AND

HIGH TECHNOLOGY PRODUCTS (PC/LAPTOP) BUYING ........................................................ 280

TABLE 9.9 RANGE AND MEANS OF PAD AND AMINUSA FOR THE EIGHT PUBLISHED STUDIES

OF THE BPM (ALL MEANS DIVIDED BY 6 SO AS TO FACILITATE COMPARISON) ........... 281

TABLE 9.10 MEANS FOR THE TWO SITUATIONS EXAMINED IN THIS STUDY. GROCERY AND

HIGH TECHNOLOGY PRODUCTS (PC/LAPTOP) BUYING. ....................................................... 281

TABLE 10.1 CORRELATION COEFFICIENTS BETWEEN KINDS OF CONFUSION IN THE TWO

MARKETS .......................................................................................................................................... 293

xiii

LIST OF FIGURES

FIGURE 1.1 EXISTING STATE OF THE UK RETAIL MARKET .............................................................. 7

FIGURE 1.2 A ROADMAP TO THIS THESIS ............................................................................................ 20

FIGURE 3.1CONSUMER CONFUSION CYCLE ........................................................................................ 41

FIGURE 3.2 INFORMATION OVERLOAD AS THE INVERTED U-CURVE.......................................... 45

FIGURE 3.3CONCEPTUALISATION OF CONFUSION IN THE MOBILE PHONE INDUSTRY .......... 48

FIGURE 3.4 CONSUMER CONFUSION TRIGGERS ................................................................................ 51

FIGURE 4.1 THE A, B, C OF CAUSATION: ANTECEDENTS (DISCRIMINATIVE STIMULI),

BEHAVIOUR AND CONSEQUENCES .............................................................................................. 70

FIGURE 4.2 THE BEHAVIOURAL PERSPECTIVE MODEL OF CONSUMER CHOICE ...................... 81

FIGURE 5.1 MEHRABIAN AND RUSSELL'S APPROACH TO ENVIRONMENTAL PSYCHOLOGY

............................................................................................................................................................. 103

FIGURE 5.2 THE BPM CONTINGENCY MATRIX ................................................................................. 111

FIGURE 6.1 BPM-E .................................................................................................................................... 125

FIGURE 6.2 BPM-E .................................................................................................................................... 126

FIGURE 6.3 BPM-I ..................................................................................................................................... 129

FIGURE 6.4 BPM-I ..................................................................................................................................... 133

FIGURE 7.1 SUMMARY OF THE RESEARCH STEPS FOR THE DEVELOPMENT OF THE MAIN

RESEARCH INSTRUMENT AND THE IMPLEMENTATION OF THIS STUDY. ........................ 167

FIGURE 7.2 DATA COLLECTION METHODS ....................................................................................... 171

FIGURE 8.1 PROCESS OF DATA ANALYSIS ........................................................................................ 210

FIGURE 8.2 SCATTERPLOTS OF THE RELATIONSHIP BETWEEN AROUSAL AND CONFUSION

............................................................................................................................................................. 240

FIGURE 8.3 NORMAL P-P PLOTS OF REGRESSION STANDARDISED RESIDUAL FOR APPROACH

............................................................................................................................................................. 253

FIGURE 8.4 SCATTERPLOTS OF STANDARDISED RESIDUALS FOR APPROACH ....................... 253

FIGURE 8.5 NORMAL P-P PLOTS OF REGRESSION STANDARDISED RESIDUAL FOR

AVOIDANCE ..................................................................................................................................... 254

FIGURE 8.6 SCATTERPLOTS OF STANDARDISED RESIDUAL FOR AVOIDANCE ....................... 254

FIGURE 8.7 NORMAL P-P PLOTS OF REGRESSION STANDARDISED RESIDUAL FOR AMINUSA

............................................................................................................................................................. 255

FIGURE 8.8 SCATTERPLOTS OF STANDARDISED RESIDUAL FOR AMINUSA ............................ 255

FIGURE 10.1 BPM-E .................................................................................................................................. 295

FIGURE 10.2 BPM-I ................................................................................................................................... 298

xiv

ABBREVIATIONS

Acronym

1 A_A Approach-Avoidance (Aminusa)

2 ANOVA Analysis of Variance

3 AP Approach behaviour

4 AV Avoidance behaviour

5 BPM Behavioural Perspective Model

6 BPM-E Behavioural Perspective Model- Extensional

7 BPM-I Behavioural Perspective Model- Intentional

8 CC Contingency Category

9 CFA Confirmatory Factor Analysis

10 CR (Classical Conditioning) Conditioned Response

11 CS (Classical Conditioning) Conditioned Stimulus

12 DV Dependent Variable

13 EFA Exploratory Factor Analysis

14 EM Expectation Maximisation

15 FA Factor Analysis

16 GLM General Linear Model

17 IGD Institute of Grocery Distribution

18 InfP Informational Punishment

19 InfR Informational Reinforcement

20 ITC Item Total Correlation

21 IV Independent Variable

22 KMO Kaiser-Meyer-Olkin (measure of sampling

adequacy)

23 MAR Missing At Random

24 MCAR Missing Completely At Random

25 MNAR Missing Not At Random

26 MR Approach Mehrabian and Russell Approach

27 MVA Missing Values Analysis

28 NPD New Product Development

29 PAD Pleasure–Arousal–Dominance

30 PCA Principal Component Analysis

31 R Response

32 RGB Rule-Governed Behaviour

33 SD

or SD Discriminative Stimulus

34 St/p

A Reinforcing or Punishing Stimulus

35 SPSS Statistical Package for the Social Sciences

(later named Statistical Product and Service

Solution)

36 TAM Technology Acceptance Model

37 TPB Theory of Planned Behaviour

38 TRA Theory of Reasoned Action

39 Tukey's HSD Tukey’s Honestly Significant Difference

40 UR (Classical Conditioning) Unconditioned Response

41 US (Classical Conditioning) Unconditioned Stimulus

42 UtilP Utilitarian Punishment

43 UtilR Utilitarian Reinforcement

44 VIF Variance Inflation Factor

CHAPTER 1- INTRODUCTION

1

1. INTRODUCTION

1.1. Introduction

This thesis investigates consumer confusion from a behavioural perspective and offers a

novel suggestion for the nature and study of the construct. It specifically describes

confusion based on the principles of rule-governed behaviour and treats it both as an

extensional consequence of behaviour and an intentional entity in two specific shopping

situations, that of grocery and high technology markets. The study uses the main

propositions put forward by the Behavioural Perspective Model (BPM) to understand

these effects and extends the model to the use of both an extensional and an intentional

language. This chapter provides an introduction to this research and begins with a

description of the research background focusing on the UK retail environment. This is

followed by the research purpose and objectives. The key research hypotheses and

justification for the conduct of this research will be presented. The subsequent part gives a

summary of the research methodology. The last sections of this chapter outline the

potential contributions for theory and practice and the structure of this thesis.

1.2. Research Background

Consumer confusion has grown into a widespread issue in retail shopping and generally in

consumer environments. It is not by chance that in the past decade (since the early 2000)

in the UK only, several comparison websites have appeared, that either directly or

indirectly imply that can help consumers encounter confusion. Price comparison websites

like www.confused.com or www.comparethemarket.com exist mainly under the

justification that these can provide consumers a simultaneous comparison of the many

different service providers (mainly insurance policies, financial services etc.) with an aim

to end confusion which derives from price offers in connection with the amount of

http://www.confused.com/

http://www.comparethemarket.com/


2

services these policies offer. At the other end non-for-profit consumer bodies like ‘which’

have been organised around product comparison websites- www.which.co.uk, publication

of magazines, product testings and offering of what has been described as ‘impartial

reviews of products’ ranging from the everyday to technology related. Their main aim as

an organisation has been to empower consumers and one of the objectives of this

consumer body refers directly to confusion and has been described as following:

‘Tackling everything from banking reform to energy tariff complexity, our commitment to

providing unbiased advice to consumers is still at the heart of everything we do...’

(Anonymous, ‘Which’ website, 2013)

It is then evident, that such a widespread concern like confusion as a business idea and

proposition can provide enough profitability to support the existence of both for profit and

non-for-profit organisations. Consumer confusion seems then to be a problem imperative

enough to justify intense research focus.

On top of the aforementioned commercial treatment of confusion several issues in the

retail environment have increased the prominence and effect of confusion in such settings.

The main point is that overall the retail industry faces a great number of issues which are

reflected to its existing state. Environmental concerns (mainly expressed in terms of

sustainability), multichannel shopping, the concept of retailtainment, the changing

consumer behaviour and the dominance of few but powerful retailers are only some of the

reasons causing an interest in retail settings during recent years and these will be briefly

analysed below (points summarised from Mintel, 2009c; Mintel, 2010a; Mintel, 2013).

1.2.1. Sustainable Retailing– Environmental Concerns

Corporate and social responsibility remains at the centre of the political and social agenda

and this movement is expected to continue especially concerning issues of sustainability,

http://www.which.co.uk/


3

environmental concerns but also responsible trading. People are now more than ever

aware that the energy and consumption intensive lifestyles have both an increased

economic and labour cost but also a possible enormous environmental impact. The agenda

of sustainability for retailers and manufacturers is long and can contain such issues like

(as in Mintel, 2009a):

ingredients and use of raw materials

the conditions those producing products work under (e.g. fair trade)

the way products are shipped and how far these are shipped

the energy costs of UK distribution networks

the energy costs and carbon footprint of the retail stores

how products are packed and what happens to waste packaging

the method used to dispose the item at the end of its life

The retail industries are in the heart of consumerism and as such have realised both their

clear responsibility to deliver environmental friendlier and ethically sourced products and

also the possible gains through a proper exploitation of this trend. New frameworks for

managing complex sustainable areas are an intimidating and costly task for most of the

retailers but these are regions that industry leaders do not ignore. General retailers and

food manufacturers have used sustainable marketing promises, fair-trade logos, reduced

packaging and carbon footprint logos extensively.

1.2.2. Multichannel Shopping on the Rise

Online retailing is growing and still gaining share of retailers’ sales. The UK online

grocery market for instance is estimated at £4.4 billion (including sales tax and delivery

charges) in 2009, having more than doubled (134% growth) in value over the period

2005–2009 (Mintel, 2009b). Notwithstanding this impressive growth, online retailing still


4

accounts for only 3% of sales of the total grocery sector, making it a niche channel when

estimated in the broader context (Mintel, 2009b). Other examples include that of the

PC/Laptop market where Mintel’s latest estimates suggest that the largest PC specialist in

the UK achieves 10% of its total sales online, equivalent to £120-140 million a year

(Mintel, 2009c).

On the consumer side, saving time, reducing shopping related stress and having control

over shopping budget have been reported as key reasons to shop online (Nielsen, 2011).

At the same time, rising broadband penetration and higher connection speeds, together

with the steady expansion of the online retailers’ geographical coverage and improved

service, has made online ordering accessible to more people.

The short term predictions for online retailing growth are mediocre and will be guided by

the limited potential for further increase in the access to online shopping, which has acted

as a main driver of sales until recently (Mintel, 2009b). However, future advances in

technology and connectivity along with consumer groups with relevant experience will

drive the future of online shopping. Mobile technology is likely to play a bigger role in the

future as more consumers are using their smart phones to help organise their lives and

shop efficiently (Nielsen, 2011). Retailer-specific mobile apps save time and provide self-

service options for consumers; tools like these are win-win strategies for both shoppers

and retailers (Nielsen, 2011).

1.2.3. Changing Consumer Behaviour

The most important implication of the recent recession is that both Mintel’s consumer

behaviour trackers and IGD back up the feeling that a new breed of shoppers is emerging

from the recession (Mintel, 2010). Both sources are reporting that altered behaviours

which embrace less and more careful spending are reported to become permanent. In


5

addition, issues of national interest and food sovereignty, the acceptability of modern

technologies, the value placed on social welfare and all that in the twilight of recession are

issues that pose a great challenge not only for UK but for the international retail system.

1.2.4. Retailtainment

Opportunities exist for retailers to take advantage of other current consumer and

marketing trends like the constant search for new experiences, particularly through the

collaboration with partner organisations (Nielsen, 2011). For example, several donut

fixtures, coffee shops and restaurants have been established within big retail stores. In

addition, food retailers add more products to their already overloaded stores and dedicate

specialised store space and aisles usually named ‘World Food’, offering more and more

delicacies from the globe. More product features are constantly added to all kind of

technology products in order to satisfy the need for change and constant improvement.

1.2.5. Market Dominance

The recent history of UK retailing has been characterised by market dominance, in the

sense that few, large companies concentrate sales, by operating larger stores (and recently

online stores) (Pal & Byrom, 2005). These larger stores have enabled the major multiple

retailers to widen product ranges, generate economies of scale, introduce and establish

their own-label brands and increase their market share. At this end increasing fears have

been addressed that the general retail environment in the UK along with the High Streets

(areas where the main retail activity in each city, town or village is traditionally taking

place in the country) have reached a level of homogenisation (‘McDonaldization’ as

described by Ritzer, 2004). Such concerns describe the way that homogenisation can

potentially lead to tiresome shopping experiences as again in this case the rules to guide

behaviour seem to be lacking distinctiveness. This state has been described as the direct


6

result of the overpowered and in many cases globalised few large retailing and

manufacturing businesses and the incapability of the independent stores or producers to

react effectively to the competition and frequently simply responding with me-too

products (Loken et al., 1986). According to Mick (2007) who takes a more general

conception of the issue, there is troubling evidence that international corporations in 21st

century life have eroded freedom of will or at least its probability. Although, the ever-

faster, never-ending race among marketers to release new products and brand extensions

should not be neglected (Schweizer, 2004), it seems that the increased levels of

homogeneity have the same effects as complexity, meaning paralysis and that both

underlie the same topic, that of confusion.

1.2.6. Overall Evaluation of the Retail Environment

Figure 1.1 acts as a challenging attempt to summarise the findings of this research

background which have taken the form of a situational exploration of the retail market.

The aim has been to place them within a framework which can facilitate an epigrammatic

but accurate understanding of the current situation in retail markets.


7

Figure 1.1 Existing state of the UK Retail Market

Source: this study; understanding developed through this market background.

Attempting a further brief description of the framework, numerous elements of the

extended macro environment operate parametrically as ‘control factors’ for the core of the

retail sector. The aim of the businesses acting in this sector is to effectively respond both

to the fundamental internal need for growth and to the consumer market trends. In order

for these goals to be achieved, diverse strategies which lead and are at the same time led

by increasing social and consumer change and market concerns, need to be adopted. All of

these diverse strategies and social conditions dictate contradictory routes of marketing

action (e.g. public policy supports less packaging but requires more on-pack information)

and create unclear environments which are characterised by the lack of market norms and

rules that can guide consumer behaviour. This general framework can stand then as a

justification for the general aim of this undertaking; meaning, to provide an explorative

Changing socio-economic conditions

(worldwide)

Shopping mechanisms

Confusion/ lack of norms that can

guide consumer behaviour

‘value for money’

convenience

Different store formats

Multichannel shopping

Retail Environment

led by few

(multinational)

players

Social TrendsMarket Growth

Building a better

future

Sustainability

(and health)

Building an attractive

shopping

environment- Variety/

Retailtainment

Less packaging/ more information on packaging/

Different kinds of products and brands (reduced

footprint/ bio etc)

Retail services/

private labels/ branding/ products

from around the world available

Changing consumer

behaviour and patterns of

shopping


8

account of consumer confusion as a way to describe the lack of rules (and in this manner

the impediment of consumer behaviour) in different retail situations.

Framework 1.1 offers then the background to the recent trends in retailing and provides

some justification to the idea that confusion is a pragmatic and central issue in retail

environments that acts to impede consumer behaviour. Retailers should perceive the act of

defining, understanding and measuring consumer confusion as an imperative task so that

they are then able to reduce it or minimise its effect if necessary. In addition, further to the

general aforementioned issue of defining and measuring confusion, very little is known on

the way confusion influences and acts along with other emotional dimensions to have an

effect on consumer behaviour. Understanding consumer confusion itself and the way it

shapes behaviour, can then provide marketers and retailers with guidance and plans to

deal with this aversive consequence of settings.

Two diverse retail settings have been used in this study in order to serve the objective of

understanding, measuring and exploring the effect of consumer confusion in retailing. The

grocery and high technology markets (focusing on the PC/Laptop buying) were chosen

as the settings for this study based on their differences on many dimensions, like the

expected levels of consumer experience, involvement, actual complexity and importance

placed on the buying decisions. The choice of these specific markets can then easily

facilitate a comparison of the markets’situational contingencies and the effect of

confusion in different settings, in the way anticipated and dictated by the main conceptual

framework of this study, the Behavioural Perspective Model.

After establishing the importance of consumer confusion for the retail markets, the

following section will adopt a more theoretical perspective. The focus will be shifted to


9

the academic literature on the topic and the attempt will be to specify the research

problem that this thesis is dealing with.

1.3. Research Problem

Both in general psychological research and in consumer behaviour literature confusion

has been a rather neglected topic of research (Schweizer, 2004). In the consumer

behaviour literature confusion has been measured and conceptualised mainly in terms of

its antecedents meaning in terms of issues like variety, ambiguity of products and

information and similarity of products (Schweizer et al., 2006; Walsh & Mitchell, 2010).

In the psychological stream of research confusion has been mainly examined as part of the

debate on the cognition-emotion distinction (Laros & Steenkamp, 2005). As part of this

debate it has been described in a number of ways ranging from a cognitive state (Storm &

Storm, 1987), a cognitive feeling (Darwin, 1872/1962; Clore, 1992), a meta-cognition

(Heiss, 2003) and possibly a pure emotion (Rozin & Cohen, 2003a).

Although this debate has been long and rigorous, an actual understanding has not been

achieved. What has been established is that there is a group of entities, confusion is part of

this group (interest and even satisfaction belong there), that possess great affective and

informational value and that more theoretical and research inquiries should specifically

focus on their study with an aim to extend their understanding (Rozin & Cohen, 2003a

and 2003b; all subsequent theoretical answers in the same volume of Emotions journal).

Although the debate on the division between emotions and cognitions has retained its

vigour, it seems not to be respected by the actual brain functions and everyday reality.

Rather, recent findings from neuroscience argue for a more ‘gradient’ approach to this

distinction; this approach implies a different relationship between relevant entities than

the actual separation between two distinct systems (Phelps, 2006; Barrett et al., 2007).


10

Accordingly, it has been seen as an imperative task to examine mixed states like

confusion from a new perspective. Specifically the one offered by furthering a noteworthy

model in understanding consumer behaviour, the Behavioural Perspective Model (BPM)

and by adopting the principles of intentional behaviourism is offering many unexploited

possibilities (Foxall, 2004; 2007b; 2013).

Moving then to the next central topic of this study, the Behavioural Perspective Model

(BPM) proposed by Foxall (1990), is an alternative approach when compared to the

dominant edifice of cognitive psychology, to understanding consumer behaviour. The

model is based on the overarching framework of radical behaviourism and the behavioural

learning tradition. The fundamental proposition of the BPM is the ‘contextual stance’,

where consumer behaviour is located at the intersection between the consumers’ learning

history and the behavioural setting. Another distinct concept of the BPM is a bifurcation

of reinforcement, which is composed of utilitarian and informational reinforcement that

are determined by consumers’ learning history and previous experiences. As a result, the

BPM proposes three formative components of consumer situations, which are: utilitarian

reinforcement, informational reinforcement, and behaviour setting scope. Rules (and rule-

governed behaviour) prescribed by social or environmental norms also have a role in this

endeavour (Foxall, 2013).

To date, most of our understanding of behavioural analysis is derived from this model and

the principles of the contextual stance (Foxall, 1999b). Although this may be sufficient to

predict and control behaviour in the laboratory, it is unable to cope and explain the

personal level of explanation, the continuity of behaviour, and the delimitation of

behaviourist interpretations (Foxall, 2007a; 2007b). Furthermore, the previous research on

the BPM has predominantly relied on the consumer situation as determined by the

extensional model and the examination of consumers’ verbal behaviour. Foxall (2007a)


11

addresses then the imperative need for the inclusion of intentionality, which can facilitate

the analysis of consumer behaviour at the personal level of explanation; however,

intentional dimensions are still mainly theoretically incorporated in the model and

basically in terms of collective intentionality (Foxall, 2013).

The present study will bridge this gap and make a case for why and how the overall

perspective proposed by the BPM can be explored based on two models, the first is

extensional (BPM-E) and the other intentional (BPM-I). The extensional model is based

on the aforementioned BPM principles and the measurement of participants’ verbal

behaviour (responses to stimuli). The intentional model is characterised by the

‘reconstruction’ of the consumer situation which in this model is described in intentional

terms (using consumers understanding and beliefs). In this context, the case is illustrated

that confusion should be represented in the field of consumer research in the retail context

by the idea that it is a case of a self-based rule (based on the principles of rule-governed

behaviour). Rules own the capacity of being described and treated using both the

extensional (as overall responses to physical and social stimuli) and the intentional

language (as rules explicable in intentional idioms) (Foxall, 2013).

As an extensional construct (BPM-E) confusion can be described in terms of verbal

behaviour and as the aversive consequence of being in retail situations that lack rules and

norms that can guide behaviour (a case of anomy, or a rule for the lack of rules-

McClosky & Schaar, 1965) and at an intentional level (BPM-I) it can take the role of a

learning history and consumer situation which implicate more of the personal

perspectives of the individuals who act in a situation. Thus the main theoretical

proposition of this thesis can be summarised in the following sentences:


12

Confusion can be described as a self-based rule. It is more specifically, a rule for the lack

of other rules (anomy). Due to its relationship with the state of affairs (environmental

situations), it can be characterised as a self-based track (in accordance with the

categorisation proposed by Zettle & Hayes, 1982) and as such it can be treated at two

levels.

At the extensional level, it can be treated as a response to specific (discriminative) stimuli

and can act along with situational contingencies to predict behaviour.

At the intentional level, it is the result of the interplay between individual perception and

specific situations and in this case it can signal consumer responses. By adopting this ‘less

scientific route’, it can be assumed that confusion can have an impact on actual situational

contingencies. Such an approach based on intentionality allows for the personal level of

explanation to be examined (see also chapter 6 for the full conceptual framework).

As a result of the special capacity of intentionality to capture the cognitive abilities of

different organisms without committing to exact hypotheses about the internal structures

that underlie their competences (Dennett, 2007) this perspective seems to offer a solution

for the understanding of the nature of confusion even within the domain of cognitive

psychology itself.

Thus the main research problem that this study is dealing with can be practically

described in terms of exploring consumer confusion in two different retail settings and

theoretically as a furthering of the concept of confusion into areas proposed by rule-

governed behaviour, testing the principles of the extensional and intentional BPM and

applying intentional behaviourism.


13

1.4. Research Purpose and Objectives

The main aim of this study is then to explore the implications of consumer confusion in

different UK retail settings. In order to implement this purpose, this study will use an

approach based on the principles of the BPM and the measurements of the Mehrabian and

Russell (1974) approach, as conceptualised and applied previously by studies of the BPM

(Foxall, 1997b and all subsequent research stream). In this manner the implications of this

study will extend beyond the original aim. The opportunity to discuss various interesting

theoretical arguments and relationships will be offered. This study has been implemented

in the UK retail environment. Beyond the obvious reasons of convenience, the basis for

this choice has been that the UK is part of the western, industrialised world with an

agreeably advanced retail environment; thus it seemed as a fine and agreeable choice for

such a study.

Based on this overall evaluation, the main objectives can be summarised by the following

sentences:

a) To explore the contextual implications of consumer confusion in two different

retail settings.

b) To provide a novel understanding of confusion based on the principles of rule-

governed behaviour and apply the theoretical propositions of a model of

consumer behaviour, the Behavioural Perspective Model (Foxall, 1990).

c) To facilitate knowledge accretion by identifying the conceptualisation of

confusion that is concordant with the aforementioned theoretical understanding

and at a more imperative level, measures the construct free from other entities

like plain frustration or annoyance.

d) To examine whether the explanation of consumer behaviour can be enhanced by

adding an aversive consequence to the original model (BPM-E) or the consumer


14

situation can be meaningfully reconstructed based on intentionality (BPM-I). In

this manner, variables will be added which extend our understanding of both the

extensional and intentional model.

e) To examine whether situational contingencies can be modified by the person’s

rule-making (personal level of explanation).

f) To find out whether the principles of the BPM can explain and adapt to

descriptions of different situations which are not manipulated as in previous

research to differ concerning their levels of reinforcement and setting scope.

1.5. Research Hypotheses

The following research hypotheses (table 1.1) have been developed to correspond and

serve the above general research objectives. These rely on previous studies of the

application of the Mehrabian and Russell (1974) measurements for the exploration of the

BPM (see Foxall, 1997b and subsequent research) and their development supports the

extended needs of this specific study.


15

Table 1.1 Research hypotheses of this study

BPM-I

H1: Overall, the range of confused consumers will indicate lower levels of Approach

behaviour than the range of non-confused consumers.

H2: Overall, the range of confused consumers will indicate higher levels of Avoidance

behaviour than the range of non-confused consumers.

H3: Overall, the range of confused consumers will indicate lower levels of Aminusa

(approach-avoidance) behaviour than the range of non-confused consumers.

H4: Overall, the range of confused consumers will indicate lower levels of Pleasure than the

range of non-confused consumers.

H5a: Overall, the range of confused consumers will indicate lower levels of Arousal than the

range of non-confused consumers (providing that consumers attribute confusion internally

and thus it has an effect on the informational reinforcement of a market).

H5b: Overall, the range of confused consumers will indicate the same levels of Arousal with

the range of non-confused consumers (providing that confusion will be attributed externally

and thus it has no effect on the information reinforcement of a market).

H6: Overall, the range of confused consumers will indicate lower levels of Dominance than

the range of non-confused consumers.

H7: The effect of confusion on Pleasure will be stronger for the market characterised by

overall lower levels of experience.

H8: The effect of confusion on Arousal will be stronger for the market characterised by


H9: The effect of confusion on Dominance will be stronger for the market characterised by


H10: The effect of confusion on Approach behaviour will be stronger for the market

characterised by overall lower levels of experience.

H11: The effect of confusion on Avoidance behaviour will be stronger for the market


H12: The effect of confusion on Aminusa (approach-avoidance) will be stronger for the

market characterised by overall lower levels of experience.

BPM-E

H13: The two markets are expected to differ in terms of utilitarian reinforcement with the

high technology market expected to have higher Pleasure than the grocery market.

H14: The two markets are expected to differ in terms of informational reinforcement with the

high technology market expected to have higher Arousal than the grocery market.

H15: The two markets are expected to differ in terms of Dominance with the high

technology market expected to have lower Dominance than the grocery market.

H16: Possible differences are expected in the levels of Confusion between the two markets

that act as discriminative stimuli in this study (a posterior or post-hoc comparison).

H17: Approach will be higher in the market characterised by higher levels of utilitarian and

informational reinforcement (thus the high technology purchasing situation).

H18: Avoidance will be higher in the market characterised by lower levels of utilitarian and

information reinforcement (thus the grocery market is expected to have higher avoidance).

H19: Aminusa, the net difference between approach and avoidance will be higher in the

market characterised by higher levels of utilitarian and informational reinforcement (thus the

high technology purchasing situation).

H20: Affective variables of Pleasure, Arousal and Dominance will each have a positive

relationship with Approach. Confusion will have a negative relationship.

H21: Affective variables of Pleasure, Arousal and Dominance will each have a negative

relationship with Avoidance. Confusion will have a positive relationship.

H22: Affective variables of Pleasure, Arousal and Dominance will each have a positive

relationship with Aminusa, the net difference between Approach and Avoidance.


16

Confusion will have a negative relationship.

H23: Aminusa (the net difference between Approach- Avoidance) will be determined by

the variables Pleasure, Arousal, Dominance and Confusion.

H24: Two-way interactions can be identified between the affective variables Pleasure,

Arousal (possibly Dominance) when examining their effect on Aminusa.

Source: this study

1.6. Justification of this Research

The conduct of this study is justifiable based on both theoretical and managerial reasons.

There are many arguments for the exploration of consumer confusion and equally many

which justify the adoption of the perspective proposed by the BPM and intentional

behaviourism. This section deals with the research justification, which is summarised in

the following bullet points:

The growing practical importance of confusion and its implications for the

practice of marketing. See for example several organisations like confused.com,

comparethemarket.com, which.co.uk and the argument that retail settings lack

clarity.

An unclear definition of the nature of confusion which has resulted in multiple

treatments both in psychology and consumer behaviour research. The necessity

to further elucidate constructs like confusion which possess both cognitive and

affective implications, has been stressed (Rozin & Cohen, 2003a).

An attempt to examine a topic with new eyes, to explore an existing landscape

from the view that a different theoretical perspective has to offer. An alternative

explanation of confusion based on theoretical arguments borrowed from

behaviourism and rule-governed behaviour can shift the focus and enhance

understanding.

A theoretical expansion of the BPM and its measurement instrument, the PAD

and approach/avoidance behaviour.


17

The examination of the fact that confusion as either an aversive consequence of

shopping or an intentional stance acts along with reinforcement and behaviour

setting scope to determine behaviour in accordance with the principles of the

BPM.

An attempt to apply, understand and further develop the principles of a recent

philosophical framework and research stream called intentional behaviourism.

In this manner, although the preceding research of the BPM has examined the

consumer situation in the extensional model solely in terms of the scope of

consumer behaviour setting (dominance), the intentional dimension of the model,

has only been rarely addressed, mainly in terms of collective intentionality.

The potential contributions of this study in terms of re-examining the way that

the original measurement of the Mehrabian and Russell model corresponds to

pure consumer/choice situations.

1.7. Research Methodology

Methodologically, this research advocates the lack of clear boundaries between social

ontologies that can allow for the mixing of approaches and the bridging of different

paradigms. Specifically, this knowledge inquiry is achieved by embracing intentional

behaviourism (Foxall, 2004; Foxall, 2007b) as its main philosophical framework, which

clearly corroborates pragmatic aspects of this study. A quantitative approach has been

used, where a main quantitative survey (N=260) has been informed by a small in extent

but meaningful in the produced results pilot- exploratory study (N=7) and multiple

(N=10) (N=56) pilot tests and discussions with knowledgeable and lay participants. The

main design of the study has been described as being explorative/ descriptive and this is in

accordance with the overall aim of this study which is the proposition of a novel

framework of understanding consumer confusion and conducting research.


18

Chapter 7 provides a detailed examination of this study’s methodology. The

methodological choices and results of this study are informed by the main theoretical

frameworks, the BPM and intentional behaviourism.

1.8. The Contributions of this Thesis

The present study will offer a novel behavioural perspective in examining consumer

confusion in retail settings. It will demonstrate the way to explore consumer confusion

based on this behavioural perspective by implementing the theoretical approach proposed

by the Behavioural Perspective Model, extending simultaneously the existing consumer

confusion and BPM literature. In this manner it adds to the existing knowledge of retailing

and consumer research and broadens the concept of confusion beyond the standard

cognitive approaches used until recently. It will also be an empirical contribution through

the indication of the use of the specific model and rule-governed behaviour in the

exploration of consumer confusion for future studies.

Furthermore, another contribution of this thesis is to introduce the role of individual

intentionality in association to the BPM. This addition offers multiple benefits; it

strengthens a) researchers’ understanding of the construct of confusion, b) researchers’

ability to analyse consumer confusion’s effects in retail settings, as well as c) the

conceptual model (BPM) itself. The empirical use of intentionality and its incorporation to

the BPM, which has been extensively discussed in the philosophy of intentional

behaviourism, is described as part of this study’s conceptual development. As a result, a

suggestion on the applicability of this philosophical framework is offered and the

extension of the relevant theoretical knowledge is achieved.

Ultimately, the findings of this study will act as a guide for marketers, companies,

retailers and policy makers especially of the grocery and the high technology markets in

the UK. The knowledge produced can be used to develop new or adjust existing product,


19

marketing or retail strategies. To start with, knowing the different dimensions of

confusion can help practitioners to improve their practices. This study might indicate that

a focus on minimising the implications of confusion might prove equally important in

improving consumers’ approach and minimising their avoidance behaviour when

compared to other environmental elements. In that sense, understanding the mechanisms

of confusion enables the marketers to tailor the marketing strategy by utilising schedules

that minimise aversive consequences (like confusion) and increase reinforcement to

control retail settings more effectively in the long run. These anticipated results will

provide an opportunity for researchers and marketing practitioners to understand

consumer confusion and then re-examine and change their existing strategies in order to

provide more appropriate and successful plans in response to this phenomenon.

1.9. Thesis Structure and Order

In order to accomplish the research objectives and hypotheses outlined in sections 1.4 and

1.5 this thesis is planned around ten chapters. Each chapter has its own unique

contribution. Specifically, chapters 2-5 which compose the literature review of this thesis

are structured and positioned in that way so that the information provided by each chapter

to add theoretical knowledge that will facilitate the overall understanding, clarify the

rationale and lead to the connections proposed in the conceptual framework (chapter 6).

Figure 1.2 offers a roadmap to this thesis:


20

Figure 1.2 A roadmap to this thesis

Source: this study

Chapter 1 provides an overview of the thesis. Thesis background, objectives and

hypotheses are described and placed into context.

Chapter 2 signals the start of the critical evaluation of the literature. The main aim of the

chapter is to summarise the treatment of confusion in psychological literature and to

indicate the requirement for further examination of interesting psychological states like

confusion.

Chapter 3 focuses on the ways that the consumer behaviour literature has dealt and

conceptualised confusion.

Chapter 4 offers the necessary knowledge background to achieve the understanding of the

theoretical model used in this study, the Behavioural Perspective Model (BPM).

Differences between behaviourism and cognitive psychology are discussed and the

Chapter 2

The Nature of Confusion in

Psychological Research

Chapter 1

Introduction and Research

Background

Chapter 5

The Mehrabian and Russell

Approach

Chapter 4

The BPM and Intentional

Behaviourism

Chapter 3

The Treatment of Confusion

in Consumer Behaviour

Chapter 6

Conceptual Framework

Chapter 7

Research Methodology

Chapter 9

Further AnalysisChapter 8

Data Analysis

Chapter 10

Discussion, Implications and

Directions for Future Research


21

chapter concludes with the imperatives of the inclusion of intentional terms in the study of

behaviour. The philosophical framework of intentional behaviourism which

accommodates both intentional and behavioural principles concludes this chapter.

Chapter 5 introduces the Mehrabian and Russell (1974) approach to consumer emotional

and behavioural responses. Previous research that has been conducted based on this

theoretical foundation is described and specifically the way these measurements have

been applied in the case of the BPM is elucidated.

Chapter 6 summarises the main points of interest of the previously examined literature

and extends the understanding of confusion based on the concept of rule-governed

behaviour. Self-based rules (in the form of tracks) are discussed as a suitable explanation

of consumer confusion and the chapter will examine the applicability and implications of

this explanation for the study of the extensional (BPM-E) and intentional (BPM-I) BPM.

Chapter 7 introduces the methodological approach of this study. Elements of the scientific

research paradigm, research design, research methods and analysis techniques utilised for

the implementation of this project will be critically reviewed. The findings of the

exploratory pilot test that preceded the main online survey are incorporated in this chapter.

Chapter 8 introduces the analysis of the main quantitative data using different statistical

tests in order to explore and answer the main hypotheses of this study.

Chapter 9 provides some further analysis in terms of consumers’ socio-demographic

characteristics and comparison of the current findings with previous studies on the BPM

and the PAD/approach-avoidance model.


22

Chapter 10 offers a theoretical discussion of the findings, considers the theoretical

suggestions of this study and the managerial implications. Finally, propositions for future

research are offered and a reflexive account on the research process is provided.

1.10. Conclusion

This chapter summarises this research by explaining the purpose and objectives, the main

justification, potential contributions and overall structure of the thesis. It has been

established that confusion is a central, contemporary market issue and according to

psychological research one of great theoretical importance. It is finally a rather mistreated

or forgotten concept in the consumer behaviour literature. On these grounds, offering a

renewed interest for the construct has been appraised as imperative. Starting with the

literature on psychological research, the next chapter will be the first on the critical

evaluation of the literature.

CHAPTER 2- LITERATURE REVIEW AND EVALUATION

23

2. THE NATURE OF CONFUSION IN

PSYCHOLOGICAL RESEARCH

2.1. Introduction

A central aim of the studies exploring the nature of confusion has been to evaluate the

entity and place it within the debate on the distinction between affective and cognitive

states. Starting on the critical evaluation of the literature of this thesis, this chapter will

commence by examining the ways confusion has been treated by previous theoretical and

research studies in psychology. This treatment will illuminate part of the reasons that an

alternative theoretical explanation for the construct will be proposed by the conceptual

framework of this study. In that sense, this chapter is an essential step before getting into

the treatment of the construct as proposed by this research.

2.2. Previous Treatments of Emotional Words

It is widely recognised that emotions do not constitute an easily definable class of entities

and thus an acceptable definition seems impossible (see Kleinginna & Kleinginna, 1981;

Fehr & Russell, 1984). Laros and Steenkamp (2005, p. 1439) attempted to combine

psychology and consumer behaviour literature with an aim to indicate that divergent

research streams can be integrated in a hierarchical model of consumer emotions. In order

to achieve this objective they have content analysed eleven seminal studies: nine from

psychology (Plutchik, 1980; Russell, 1980; Watson & Tellegen, 1985; Shaver et al., 1987;

Storm & Storm, 1987; Morgan & Heise, 1988; Watson et al., 1988; Frijda et al., 1989;

Roseman et al., 1996) and two from consumer behaviour (Havlena et al., 1989; Richins,

1997) and have compiled a comprehensive table of 143 positive and 173 negative emotion

words (refer to table 2.1, next page).


24

Richins (1997) was the first to signify that the general emotional measures, widely

adopted by her contemporary flourishing consumer literature on emotions, could not

adequately depict the emotional implications of consumption-related situations. On these

grounds, a specific Consumption Emotion Set was constructed. These consumption

emotion words are depicted in italics in the table.

Table 2.1 173 negative and 143 positive emotion words

Negative Emotion Words Positive Emotion Words

Aggravation a,b,c, Agitation a,b,c, Agony b,c,

Alarm b,c,d, Alienation b, Anger a,b,c,d,e,f,g,

Anguish a,b,c, Annoyance a,b,c,d,e,f,h, Anxiety

a,b,c,e, Apologetic c, Apprehension a,b,c, Aversion

e, Awful c, Bad c, Bashful c, Betrayal c, Bitterness

a,b,c, Blue a,c,i, Bothered c, Cheerless a, Confused

c, h, Consternation c, Contempt b,c,e,g, Cranky c,

Cross c, Crushed h, Cry c, Defeat b, Deflated a,b,

Defensive c, Dejection a,b,c, Demoralized c,

Depression a,b,c,d,h, Despair b,c, Devastation c,

Different c, Disappointment a,b,c,e,f, Discomfort c,

Discontent a,c, Discouraged c, Disenchantment c,

Disgust a,b,c,e,g,h, Dislike b,c,g, Dismay b,c,

Displeasure a,b,c, Dissatisfied a,c, Distress

a,b,c,d,g,i,j, Distrust c,e, Disturbed c, Down a,c,

Dread b,c, Dumb c, Edgy c, Embarrassment a,b,c,

Empty a,c, Envy a,b,c, Exasperation b, Fear

b,c,d,e,f,g,h,i,j, Fed-up a, Ferocity b, Flustered a,

Forlorn c, Foolish c, Frantic c, Fright a,b,c,h,

Frustration a,b,c,d,f,g, Fury a,b,c, Gloom b,c,d,h,

Glumness b, Grief a,b,c,f, Grouchiness b,c,i,

Grumpiness b,c,i, Guilt b,c,e,g,j, Heart-broken a,c,

Hate b,c,Hollow c, Homesickness a,b,c,

Hopelessness b,c, Horrible c, Horror a,b,c,f,

Hostility b,c,h,i,j, Humiliation b,c, Hurt a,b,c,

Hysteria b, Impatient a,c, Indignant c, Inferior c,

Insecurity b, Insult b,c, Intimidated h, Irate a,c,

Irked a, Irritation a,b,c,h,j, Isolation b,c, Jealousy

a,b,c,e, Jittery i,j, Joyless a, Jumpy c, Loathing b,

Loneliness a,b,c,i, Longing c, Loss c, Lovesick a,

Low a,c, Mad a,c, Melancholy b,c, Misery a,b,c,d,

Misunderstood c, Moping c, Mortification a,b,

Mournful c, Neglect b,c, Nervousness a,b,c,i,j,

Nostalgia c, Offended h, Oppressed c, Outrage

Acceptance c,h, Accomplished c, Active i,j,

Admiration c, Adoration b,c, Affection b,c,

Agreement c, Alert h,j, Amazement b,

Amusement a,b,c, Anticipation b,c,

Appreciation c, Ardent c, Arousal a,b,d,

Astonishment b,d,i, At ease a,d, Attentive h,j,

Attraction b,c, Avid c, Bliss b, Brave c, Calm

a,d, Caring b,c, Charmed a, Cheerfulness

a,b,c,h, Comfortable c, Compassion b,c,

Considerate c, Concern c, Contentment

a,b,c,d,i, Courageous c, Curious h, Delight

a,b,c,d,h, Desire b,c, Determined j, Devotion

c, Eagerness b,c, Ecstasy a,b,c, Elation

a,b,c,i, Empathy c, Enchanted c,

Encouraging c, Energetic f, Enjoyment b,c,f,

Entertained c, Enthrallment b, Enthusiasm

b,c,e,f,i,j, Euphoria b,c, Excellent c,

Excitement a,b,c,d,f,i,j, Exhilaration b,f

Expectant c, Exuberant c, Fantastic c,

Fascinated e, Fine c, Fondness b,c,

Forgiving c, Friendly c, Fulfilment c, Gaiety

b,c, Generous c, Giggly c, Giving c,

Gladness a,b,c,d, Glee b,c, Good c, Gratitude

c, Great c, Happiness a,b,c,d,e,f,h,i,

Harmony c, Helpful c,h, High c, Hope b,c,g,

Horny c, Impressed c, Incredible c,

Infatuation b,c, Inspired j, Interested f,j,

Jolliness b, Joviality b, Joy a,b,c,e,f,g,

Jubilation b,c, Kindly c,i, Lighthearted c,

Liking b,c,g, Longing b, Love a,b,c,e, Lust

b,c, Merriment c, Moved a, Nice c, Optimism

b, Overjoyed a,c, Passion a,b,c, Peaceful c,f,

Peppy i, Perfect c, Pity c, Playful c, Pleasure

a,c,d,f,i, Pride a,b,c,e,f,g,j, Protective c,


25

a,b,c, Overwhelmed a, Pain c, Panic b,c, Petrified

a,c, Pity a,b,c, Puzzled h, Rage b,c,e, Regret

a,b,c,e,g, Rejection b,c, Remorse a,b,c,

Reproachful c, Resentment a,b,c, Revulsion b,

Ridiculous c, Rotten c, Sadness a,b,c,d,e,f,g,h,i,

Scared a,c,h,j, Scorn b,c,i, Self-conscious c, Shame

a,b,c,e,g,j, Sheepish c, Shock a,b,c, Shy c, Sickened

a,c, Small c, Sorrow a,b,c,e,i, Spite b, Startled e,h,

Strained c, Stupid c, Subdued c, Suffering b,c,

Suspense c, Sympathy b, Tenseness b,c,h, Terrible

c, Terror a,b,c, Threatened h, Torment a,b,c,

Troubled c, Tremulous c, Ugly c, Uneasiness a,b,c,

Unfulfilled, Unhappiness a,b,c,i, Unpleasant h,

Unsatisfied c, Unwanted c, Upset a,c,e,j,

Vengefulness b,c, Want c, Wistful c, Woe b,c,

Worry b,c, Wrath b,c, Yearning c

Rapture b, Reassured c, Regard c, Rejoice c,

Relaxed c,d,f, Release c, Relief a,b,c,e,f,g,

Respect c, Reverence c, Romantic c,

Satisfaction a,b,c,d,f,i, Secure c, Sensational

c, Sensitive c, Sensual c, Sentimentality b,c,

Serene d,c, Sexy c, Sincere c, Strong i,j,

Super c, Surprise b,e,f,i, Tenderness b,c,

Terrific c, Thoughtful c, Thrill a,b,c,

Touched a, Tranquillity c, Triumph b, Trust

c,h, Victorious c, Warm-hearted c,i,

Wonderful c, Worship c, Zeal b, Zest b

Note: The emotion words of Richins’ CES, Consumption Emotion Set (1997) are in italics.

a Morgan & Heise (1988)

b Shaver et al. (1987)

c Storm & Storm (1987)

d Russell (1980)

e Frijda et al. (1989)

f Havlena et al. (1989)

g Roseman et al. (1996)

h Plutchik (1980)

i Watson & Tellegen (1985)

j Watson et al. (1988)

Source: Laros and Steenkamp (2005, p. 1439).

This lack of consensus on emotions is depicted very accurately on table 2.1. There is

considerable agreement that certain states (like anger, fear and happiness) should be

considered emotions, and that certain others should not (hunger, thirst). There are,

however, other states which are followed by little consensus (nostalgia, dismay, interest or

confusion are just some of these states). Such states have been treated with considerable

variation in research and their definition and categorisation very much depends on the

adopted approach or research interests of each researcher. This is why the debate on the

distinction among cognitive, emotional and volitional states has resulted to both extensive

intellectual debates and meticulous literature (Frijda, 2008, p. 70).


26

Recent findings from neuroscience (LeDoux, 1996; Phelps 2006; Barrett et al., 2007)

however support the notion that the ‘distinction between cognitive activity and emotion

experience is probably better conceptualized as more of a gradient (in the sense of a

graded change) rather than two independent systems that can interact with one another’

(Barrett et al., 2007). Neuroscience indicates that the parts of the brain responsible for

emotions are the ones responsible for cognitive activity (Foxall, 2011) and thus the

separation of cognition and affect as explained and examined by psychologists is not

really respected by afferent/efferent processes. Searching for such distinctions is then

undesirable (‘Descartes error’ according to Damasio (2005) was the dualistic separation of

mind and body, rationality and emotion, feeling and thinking). Nevertheless the debate

maintains its rigor and will be exemplified here with a special interest on the state of

confusion.

2.3. Separating Emotions from Other Entities

It seems then that a crucial task to start with would be to separate emotions from

cognitions and secondly from other sensational and motivational states. Nonetheless this

undertaking requires intense effort (Bennett & Hacker, 2003) because the debate

regarding the relationship and particularly the causal role of each in the formation of the

other has been intense (Zajonc, 1980; Lazarus, 1984; Plutchik, 1985; Frijda, 2008, p. 70).

Hacker (2004, p. 200–201; also Bennett & Hacker, 2005) distinguishes among affections

(including emotions and moods), appetites, sensations and tactile perceptions. Smith &

Lazarus (1992) discern emotions from other entities that serve adaptive and motivational

functions like reflexes (e.g. startle) and physiological drives (e.g. hunger, thirst) and

Bagozzi et al., (1999) point out the difference among affect (treated as an umbrella term),

emotions (states which are situational and directed towards intentional objects), moods


27

(e.g. persistent, enduring anger; see also Gardner, 1985) and (possibly) emotional

attitudes. Scherer (2005; see also Smith & Lazarus, 1992; Zelenski & Larsen, 2000;

Frijda, 2008) further identifies that many stable character traits and behavioural tendencies

have an intense affective core (a person can be characterised as nervous, anxious, hostile,

jealous or being afraid of dogs). Such affective traits indicate that a person might have a

tendency to react in a certain emotional pattern under many diverse life circumstances and

the slightest of provocation. On these grounds, emotions (in the same way as cognitive

characteristics) can be studied both as a general personality trait and as a state generated

by particular encounters with the environment; in that second case the state does not

necessarily represent recurrent problems.

A closer examination of the relevant literature revealed three potential issues that could be

the cause of such polyphony regarding what constitutes or not an emotion.

2.3.1. The Classical View of a Concept

Firstly, an interpretation of the issue was brought to the fore by Fehr & Russell (1984).

Their theoretical explanation expands on the shortcomings of the classical view of a

concept as ‘mental pigeon holes with clear boundaries’ (p. 465). They argue that such

treatment could provide accurate descriptions for almost any everyday concept (like for

example a table as part of home furniture) but cannot be accurately applied on

psychological entities like personality or emotions. They rather support a prototypical

approach of categorisation (Rosch, 1975 as in Fehr & Russell, 1984; Shaver et al., 1987).

Such an approach advocates that concepts are better organised loosely around their

clearest examples, referred as prototypes. Following seven exploratory seminal studies

they found support for the notion that rather than trying to identify the ultimate and clear

definition of emotions a much more fluid perspective can provide better insights.


28

Since participants of the studies were found to be able to clearly indicate which emotions

are better and which are worst examples of the category, a term should be categorised as

‘emotion’ based on the ways it can be placed around ‘emotion prototypes’ rather than

based on a strict positioning in or out of the emotion category.

2.3.2. Everyday Use of Language

Secondly (and unquestionably connected to the third point below), some problems

originate from the everyday use of language by lay people (Smith & Lazarus, 1992). In

this context, a primary concern is the linguistic properties of the English words used to

define and accompany emotions. For example, the use of the word ‘feel’ in all kind of

situations—cognitive, emotional, physiological—complicates the distinction among states

(as in both I feel and I am hungry, scared, connected, satisfied, confused) and has been a

major research issue due to the fact that researchers’ typologies of what belongs in each

category seems not to be respected by every-day language and experience.

2.3.3. Criteria for Measuring Emotions

Thirdly, again according to Smith & Lazarus (1992), issues derive from the diverse

‘defining criteria’ and measurement of emotions. The incidents revealing emotions that

scientists can actually measure are traditionally threefold (e.g., Lang et al., 1998): (1) the

language of emotion (expressive and evaluative); (2) reflexive physiological activity

(somatic and autonomic reactions such as characteristic facial expressions (Ekman &

Friezen, 1971; Izard, 1994; Ekman, 2001; 2003); and (3) behavioural (e.g., approach and

avoidance, ‘freezing’, and performance deficits or enhancements). The issue remains that

there are no well accepted defining criteria and the measurement of emotion

manifestations have not been enough to provide an acceptable categorisation of emotions

until now.


29

To set some examples on this issue, the language of emotions can be used in both the ‘I

am’ and ‘I feel’ (I am happy and I feel happy) context and several human facial

expressions (see also p. 45 onwards), can be the manifestations of states that have a

cognitive rather than an affective core, like scepticism.

In order to deal with these issues Ortony and Clore were among the first to conduct a

series of studies to determine the actual domain of pure emotion terms, relatively free of

trait, physical or cognitive implications (Clore, Ortony & Foss, 1987; Ortony, Clore &

Foss, 1987). Each of their studies employs judgements of the appropriateness of words as

descriptors of emotions. The different studies vary both the samples of judges and the

linguistic frames for judgements. Their main theory is that the dimensional structure of the

lexicon of pure emotion adjectives are those which samples of lay and expert judges rate

as emotions in both the ‘feeling __’ and the ‘being __’ frames (Clore et al., 1987). All

their studies reach similar conclusions concerning the domain of emotions: the categories

created and which can be easily discriminated were four classes of entities consisting of

affective, cognitive, external and bodily conditions. The list of adjectives dealing with

internal, mental feeling states and whose focus is solely on affect consists of only about

one-quarter of the 500 words used in previous studies (as for example in Averill, 1975).

Following this brief introduction on the categorisation and nature of emotions and

cognitions, the rest of this chapter will focus on the state of confusion. As explained the

nature of some states is ambiguous and theoretical and empirical efforts have focused

until now in understanding the nature of such entities. Confusion is one of these states that

have been characterised as either emotional and/or cognitive, as a personality trait or a

response to specific stimuli. The rest of this chapter will exemplify the ways the literature

has dealt with this issue.


30

2.4. The State of Confusion

As previously explained, the term confusion can be considered a problematic one due to

the difficulty of recognising and categorising what kind of state it constitutes. It is

common to hear lay people unconsciously positioning the term in equally the ‘feeling’ and

‘being’ frames as in both ‘I feel confused’ and ‘I am confused’. The following sections

will describe the different propositions on the nature of confusion.

2.4.1. Confusion as a Cognitive State

Confusion has been particularly investigated in clinical (Banister, 2000) and gerontology

studies (Slater & Lipman, 1977; Neelon et al., 1996) to describe such situations where

individuals are unsure about how to interpret certain stimuli and act accordingly. It is then

the research norm to perceive confusion as a cognitive state, when, as Keltner and Shiota

(2003) point out, one can be overloaded with information and uncertain about what to do

or how to act.

Storm and Storm (1987) attempted to develop a hierarchical model of semantically

homogenous groups of emotional terms along with the features defining each group and

explore their central characteristics. The research was conducted in two phases. In the first

phase, hierarchical clustering of 72 terms, considered part of the emotion domain, was

used to identify a preliminary organisation. Subjects were instructed to sort the 72 terms

into non-overlapping groups according to similarity of meaning. In the second phase, four

highly educated speakers of English, sensitive to fine verbal distinctions, agreed on the

classification of a larger set of terms, using the groups identified in the first phase as a

starting point. The final taxonomy offers seven categories of emotional language: 1)

negative terms related to shame, sadness and pain, 2) negative terms related to anxiety and

fear, 3) negative terms related to hatred and disgust, 4) positive terms related to


31

interpersonal reference, 5) positive terms without interpersonal reference, 6) terms related

to activity, passivity and finally 7) cognitive states. Confusion, baffled, bewilderment,

helpless, mixed up, perplexed, puzzled and topsy-turvy are part of the cognitive category

connected to ‘a lack of control in a cognitive sense’ as the researchers explain.

2.4.2. Confusion as a Cognitive Feeling

The categorisation of confusion as a cognitive feeling follows Clore (1992, p. 141) who

specifically cited confusion as an example of a ‘cognitive feeling’, that is a feeling about

our state of knowledge. This description very much connects confusion to a state of not

knowing, which like the state of knowing is a feeling state, mainly associated with a meta-

cognition (Hess, 2003), a cognition about cognition or knowing about knowing. Clore

(1992) explains that saying that these states are not affective does not mean that they

cannot be the cause of affective or emotional reactions. As explained by this author we

may be happy that we are certain about something or frustrated that we are confused, but

being certain or confused are not themselves emotional feelings (Clore et. al, 1987).

Rather, confusion between intense cognitive states and emotions can arise because intense

cognitive states are likely to cause emotions (Clore et al., 1987).

The category of ‘cognitive feelings’ has also been explained as a certain set of feelings

involving a combination of both emotions and thought. Attention must be brought to the

fact that this explanation emphasises an intertwined approach and is not the same with the

previous explanation of a ‘cognitive feeling’ which acts as a meta-cognitive thought

giving rise/causing emotions. This latter intertwined description is more in accordance

with a set of states Darwin deals with in the ninth chapter of his 1872/1962 book, The

Expression of the Emotions in Man and Animals entitled ‘Reflection – Meditation – Ill-

Temper – Sulkiness – Determination’. There he deals with a similar kind of emotions like

perplexed reflection, which is synonymous with confusion (terms have been described as


32

synonymous by Ellsworth, 2003 and thus such an observation is not only the

interpretation of this research). Darwin (1872/1962) refers to such emotions as

intellectual, and he thought that because these emotions include both cognition and affect,

this was a good enough reason to exclude them from the words used to describe emotions.

2.4.3. Confusion as an Emotion

Drawing once again on table 2.1, p. 36-37 (developed by Laros & Steenkamp, 2005)

confusion has been attributed the qualities of either a pure or a basic emotion only rarely.

A pure emotion is any state which is free from other entities like cognition and in that

sense it has only an affective core. On similar grounds the theory behind basic emotions

maintains that emotions are distinct and measurable entities that differ from one another in

important ways.

Confusion is clearly not included in Plutchik’s well known inventory of eight basic

emotions but it has found a role in his general emotion theory (Plutchik, 1980; 1994).

Confusion found its angular location on a circle of emotions (the circle was constructed

based upon similarity of the terms) and it was positioned very close to words like

hopeless, depressed, unhappy, disappointed, uncertain, bewildered, perplexed and finally

surprised (Plutchik, 1994, p. 70).

2.5. Re-examining the Debate on Confusion

A notable study which initiated an intense intellectual discussion on the topic of confusion

and other similar states was published in 2003. Rozin and Cohen (2003a) conducted an

introductory psychology class observation study. The study was carried out by students

(n=255) and the aim was to test the valence lateralisation hypothesis (the issue concerns

the lateralisation of brain function in emotion, see Rozin & Cohen, 2003a for more

details). As part of the study, students were asked to record up to five instances of


33

spontaneous, asymmetric, facial expressions (facial actions that last seconds) and to

indicate among other things, a) the emotion expressed according to students’ judgement

(referring to whatever was going on when the expression was occurring) and when

possible b) the emotion reported by the participant. The results were based on over 2,000

observations (1,245 symmetric and 996 asymmetric facial expressions1) and extended

further from the initial goal of examining lateralisation and valence.

Surprisingly, the most interesting finding of the study is that many facial expressions

reported do not correspond to the so called basic emotions or even anything usually called

an emotion. The most striking ‘non-standard’ emotion was found to be confusion.

Confusion scored highest of all asymmetric emotions (14%) and seventh (5%) among the

symmetric emotions.

According to the authors this finding suggests that either facial expressions a golden

standard in the research of emotions, are used to a great extend to express other things

than emotions or that the category of emotions should be expanded to include more states

than the existing. The authors acknowledge that facial expressions can exist independently

of emotions and can well accompany other non emotional forms of expression like

scepticism and disbelief (Ekman, 1978; 1979 as in Cohen and Rozin, 2003a); yet they

believe that confusion would qualify as an emotion according to the criteria set out by

Ekman (1992). Confusion is surely a valenced state (negative); it has a distinct facial

expression (involving the eyes and eyebrows) and a distinct internal state (or ‘qualia’).

The study received a flurry of responses (Ellsworth, 2003; Hess, 2003; Keltner & Shiota,

2003; Rozin & Cohen, 2003b) in the same volume of the ‘Emotion’ journal. The study

1 Symmetric facial expression: where both sides of the face are equally expressive. Asymmetric: where one

side of the face is more expressive than the other (as in Rozin & Cohen, 2003a).


34

was criticised on the grounds of its biased context as confusion is highly likely to occur

when untrained college students are sent out to find examples of facial expressions. In

addition, according to most of the commentators, expressions like the ones singled by

Rozin and Cohen (2003) have been described variously as a sign that someone has

encountered an obstacle or experiences motivational incongruence (Hess, 2003), without

necessarily placing such entities in a specific category of states like emotions.

Ellsworth (2003) explains how positioning states into rigid categories like emotions and

cognitions emphasises the certainty of psychological theory over the fluid reality of

human experience. She explains how appraisal theories of emotions can accommodate

such ambiguous emotional states and describes confusion as the experience of a

combination of high level of attention and perceived effort combined with uncertainty,

helplessness and possibly goal obstruction added to the mix. As the situation changes,

appraisals also change the emotional state from one to the other or even to some unnamed

state in between, explaining why lay people easily mix up diverse description of situations

into something else. In the multidimensional space defined by appraisal dimensions,

Ellsworth positions confusion (like Plutchik, 1994) very close to surprise.

The findings of the study by Rozin and Cohen (2003) might not be conclusive but on the

grounds of these propositions, Rozin and Cohen (2003a and b) and all commentaries that

followed their publication call for more attention to some relatively ignored but common

states that have both affective content and important informational value, like

confusion. Most contemporary studies and especially studies from neuroscience support

the notion that affect plays a major role in cognition, and cognition in affect, and

acknowledge that states that clearly involve both reason and passion, offer an exciting

opportunity to study more diverse emotional experiences.


35

2.6. Conclusion

This chapter has described the ways recent psychological research has dealt with states

which hold both ‘affective and informational value’. These entities have resulted to

intense theoretical debates and for many years have been placed in the affect, the

cognition or a mixed group of states. The requirement to further elucidate and understand

such terms, to reveal their multiple characters and discuss their role in theoretical and

empirical undertakings has been stressed by previous researchers (Rozin & Cohen, 2003a;

2003b). The emphasis of this chapter has been placed on confusion which has been

described as one of these states.

The main thesis of this study is based on a behavioural perspective in research and has

been based on a proposition that confusion can be faced as a self-based rule. Such terms

have two avenues for examination (Foxall, 2013): the first is in extensional terms and the

other in intentional terms. Both of these examinations and especially the ‘intentional

nature’ (based on the concept of intentionality) can have multiple benefits for the study of

confusion. Such a conception can shed new light to its nature and based on the properties

of rule-governed behaviour can give new avenues for its examination. This

conceptualisation can further provide a way towards rejecting unnecessary dualistic

conceptions of entities (the distinction between cognition and affect). Such dualistic

conceptions are not in accordance with recent findings from neuroscience. These issues

will be properly deployed in the conceptual framework chapter but before extending even

further on these issues the next chapter will deal with the conceptualisation and treatment

of confusion in the consumer behaviour literature.


36

3. THE TREATMENT OF CONFUSION IN

CONSUMER BEHAVIOUR LITERATURE

3.1. Introduction

An ambiguous but well-known quote (most often attributed to the US president Harry S.

Truman) proposes an alternative technique of effective persuasion: ‘If you cannot

convince them, confuse them’. Several research papers in the consumer behaviour realm

have been produced in an attempt to investigate whether the business world has put this

proposition into force. Following the exploration of the ways general psychological

research (see previous chapter) has dealt with confusion as a special, however ambiguous

(concerning its nature) state, this chapter will examine and debate whether consumers

actually have the capacity, motivation and opportunity to take advantage of the freedom to

choose ‘extraordinary’, novel and abundant experiences and products, explicitly offered

by the industry (Poiesz, 2004).

3.2. Seeking Variety and ‘Extraordinary Experiences’

Abrahams (1986, p. 59) claims that especially the individualistic western societies place a

distinctive emphasis on the concept of experience and especially the notions of novelty

and variety-seeking, accompanied by an intense fear of boredom. Anyhow lay wisdom

appraises that variety can be the ‘spice of life’. In the same spirit, Ratner & Kahn (2002)

have showed that even our impressions of others are influenced by this constant search for

diversity, pointing to the fact that social motivation and self-presentation benefits often

contribute to the anticipation that restricting one’s choices to favourite items only, might

make a negative impression on others.


37

A flourishing consumer behaviour literature, particularly fervent over the recent past

decades, has further advocated the concept of variety-seeking behaviour (Kahn & Ratner,

2005). Faced either as a necessary function for growth and adaptation to a changing

environment (Foxall, 1993), an individual need for stimulation (Steenkamp &

Baumgartner, 1992) or exploratory behaviour (Berlyne, 1960), a desire to overcome

satiation (McAllister, 1982) or to identify novel stimuli (Faison, 1977) or simply as a

‘hedge’ against uncertainty regarding future preferences (Kahn, 1995; Chernev, 2006)

variety-seeking behaviour has been the topic of intense research.

In order to respond to this intrinsic or extrinsic search for variety, markets have adopted

several courses of action. First of all, high-variety strategies (mainly concerning product

and promotion proliferation) have been widely presented as a way that a company can

gain competitive advantage (Kahn, 1998). The adaptation of such strategies offer

consumers the desired numerous choices and according to some research are highly

valued by them (Oppewal & Koelemeijer, 2005). Retail settings, well tied with the

ideology of consumption, offer elements like a cornucopia of products accompanied by

objective facts (i.e. product information like nutritional facts) but also symbolic and

hedonic aspects (Levy, 1959; Hirschman & Holbrook, 1982; Belk, 1988)—which can well

be summarised as ‘the power of branding and positioning when it comes to product

differentiation’ (Levitt, 1980). Secondly, several marketing concepts and approaches, with

most recent the ‘experiential’ marketing approach, have been introduced through the years

proposing a novel marketing rationale and further opportunities for differentiation.

From classic economic and choice theories (see Lancaster, 1990 for a review) to

psychological theory and research, a link has been theoretically and empirically

demonstrated to exist between the provision of choice (and thus of personal freedom) and

increases in intrinsic motivation, task performance, perceived control, the sense of self-


38

determination, psychological well-being and life satisfaction (Rotter, 1966; Brehm, 1972;

Deci, 1975; Zuckerman et al., 1978; Deci & Ryan, 1985; Taylor & Brown, 1988). These

psychological benefits can be well connected to more practical aspects of larger

assortments which are well documented in consumer behaviour literature (Kahn, 1998;

Broniarczyk, 2008, p. 759). Given the heterogeneity in consumers’ tastes and needs, high

variability increases the ability to find the ideal (in some cases personalised) alternative; it

satisfies consumer variety-seeking behaviour; it offers flexibility for uncertain preferences

especially when future choices are negotiated and opportunity to accommodate the

requirements of multiple users. Finally, it can fulfil other non-purchase reasons of

shopping, like enjoying shopping as a leisure activity and trend watching (Bellenger &

Korgaonkar, 1990; Guiry et al., 2006).

It has also been asserted that given the fact that consumers’ are presented with the ability

to exercise free choice, a shift of social and market power from the producers to the

consumer has taken place (Denegri-Knott et al., 2006). This increased freedom to choose

is often associated with a higher living standard and greater consumer power, also

characterised as ‘market or consumer democracy’ (Lane, 2000; Schweizer et al., 2006).

This notion of freedom is also embraced by some approaches based on the eclectic school

of existentialism. On these grounds, over-choice has been explained as a consumer’s right

to exercise freedom and to create new meanings, for the plethora of goods, through one’s

own idiosyncratic self (Elliot & Ritson, 1995; Baumeister et al., 2008). In that sense,

freedom equals to being able to express oneself and to take pleasure through the

experience of buying and using products (Cross, 2000).

On the same foundations with choice and variety, the marketing concept itself has moved

from the introduction of the ‘4 Ps’ (McCarthy, 1964) and the ‘transaction or exchange’

notion (Bagozzi, 1974) to the firm and innovative concept of ‘relationship marketing’ in


39

the 90’s (Dwyer et al., 1987; Morgan & Hunt, 1994) and has been attempting to find its

way to a new era, linked to the concept of ‘experience’ (Holbrook & Hirschman, 1982;

Schmitt, 1999; Carù & Cova, 2003). Although the suggestion that all consumer

experiences are (or ought to be) ‘extraordinary’ has been dealt with increasing criticism

(see Carù & Cova, 2003), marketers and marketing academics support the notion that the

continuous search for novel and meaningful experiences, as a fundamental need of the

‘millennial consumer’ (Holbrook, 2000), should be satisfied during every consumption or

purchase occasion by the engagement of all the human senses. Such engagement allows

the formation of a connection between the consumer and the product/service, on both

emotional and rational grounds. The provision of ‘experiences’ rather than products or

services has been proposed as a solution for any kind of business including pure retailing

(Kim, 2001), where the aim is the creation of retailtainment or shoptainment. The ultimate

motive there is to offer consumers physical and emotional sensations and richness before,

while and after the shopping itself.

3.3. Some Opposing Arguments

However, this supposition that more complexity is necessarily beneficial has been

increasingly questioned by many groups of researchers from different disciplines. To start

with, no matter how valued variety-seeking is, it is documented that the actual need for

variety differs by individual (Steenkamp & Baumgartner, 1992). Especially, the concept

of Optimum Stimulation Level (OSL) (Raju, 1980) specifies that there is heterogeneity

with regard to how much stimulation an individual may feel as optimal and desirable.

Depending on their stimulation level consumers might feel in unease with arousal which

is either above or under their stimulation acceptance level.


40

Iyengar & Lepper (2000) argue that although the argument for the psychological benefits

of choice seems compelling, the number of choices/alternatives presented in relevant

experiments is usually small ranging between two and six alternatives. This evidently

indicates that relatively limited choice among alternatives is more beneficial than no

choice at all. However, their results point to the fact that people actually seem to prefer to

exercise their opportunity to choose in contexts where their choices are limited, and they

even perform better in such limited-choice contexts. The question then occurs what

happens in real life shopping situations—which are much more complex than few

alternatives—where consumers are also faced with limited time? A growing body of

literature argues for the existence of the phenomenon of the ‘tyranny of choice’ as Barry

Schwartz (2004) named the phenomenon of increasing consumer choice resulting to

suffering. The most influential stream of research in this tradition is the one dealing with

assortment size indicating that consumers face more difficulty to choose from larger than

smaller assortments (Iyengar & Lepper, 2000; Schwartz, 2000; 2004; Chernev, 2003a &

b; Fasolo et al., 2006). These recent findings nevertheless reflect and build upon older and

well-known research findings and discussion of the 70’s and 80’s (Jacoby et al., 1974;

Malhotra, 1982 and all subsequent flurry of responses) which had already established that

more choice can be less.

At the more practical level of business reality, the benefits (along with the shortcomings)

of the aforementioned strategies bear implications at the level of both marketing and

strategy practice (Bayus & Putsis, 1999). The confusion cycle (figure 3.3) as proposed by

Schweizer (2004) identifies that the source of ‘pluralistic’ business strategies lies with the

consumer and is attributed to the fact that consumers no more follow discrete, logical

shopping rules, rendering the identification of buying patterns into a difficult task. On the

consumer side however, these strategies cause confusion. In this manner, while consumers


41

intent to maintain the overview of the busy environment they employ confusion reduction

strategies causing more difficult identification of the buying patterns and in that way

closing the confusion cycle (see figure 3.1 below).

Figure 3.1Consumer confusion cycle

Source: Schweizer, 2004, p. 3 (translated into English by the author).

3.4. Consumer Confusion

Consumer confusion is an existing problem in contemporary markets and this has been

identified in many market situations such as computer software and multi-media

(Khermouch, 1994; Cahill, 1995), telecommunications (Turnbull, et al., 2000; Leek &

Chansawatkit, 2006), insurance and mortgage markets (Woodward & Hall, 2010), watch

market (Mitchell & Papavassiliou, 1997), food labelling and beliefs about diet and food

(Golodner, 1993; Ippolito & Mathios, 1994; Marshall, et al., 1994; Wiseman, 1994; West

at al., 2002; Hasler, 2008; Abrams et al., 2010; Henryks & Pearson, 2010), higher

education (Drummond, 2004), recycling symbols and environmentally-friendly claims

(Mendleson & Polonsky, 1995), wine market (Drummond & Rule, 2005; Casini et al.,

Management Perspective

Consumer Perspective

Reduced Sales

Shopping Fatigue

More difficult identification of buying patterns

PluralisticStrategy

Consumer Confusion

Reduction Strategies


42

2008), general grocery shopping—as early as 1966—(Friedman, 1966), complaint

channels in public services (Ashton, 1993), advertisement (Elliott & Speck, 1998) and on

the internet (Walsh et al., 2004).

3.4.1. Initial Conceptions of Confusion

Consumer confusion itself might have been a neglected phenomenon in the books of

consumer behaviour, nevertheless confusion triggers have been cited, mentioned and

investigated to a great extend but either at an isolated manner or because these serve or

disturb other aspects of consumer behaviour like consumer attitudes, choice and decision-

making. However, examining a chronological review and synopsis on the confusion

related literature (as in Appendix 1, this study), the problem arises that much of the

aforementioned published material is either opinion articles and market reports (e.g.

Boxer & Lloydd, 1994; Drummond, 2004) or research studies which employ a very broad

conception of confusion (e.g. West et al., 2002; Casini et al, 2008). In that case, consumer

confusion operates as a general facilitative framework that allows researchers to skip any

attempt to conceptualise or comprehend the confusion construct itself. The idea of

confusion then acts as evidence for the inevitability of the main study, which most of the

times is linked to either the investigation of consumer attitudes or the identification of

possible environmental causes of consumer confusion in a specific market. An illustration

of the first case can be extracted from the study by West et al. (2002) who employed the

concept of confusion in the veal market in order to justify the main objective of their

study, to investigate consumer preferences toward veal attributes. At the other end, Casini

et al. (2008) used consumer confusion as a framework in order to test the existence of the

principal elements of consumer misunderstanding in the buying process of wine. Using

the method of the ‘key informant’ as an expert source of information the study identified

the causes of consumer confusion in the wine market to be: availability of access to


43

increasing amounts of information, product proliferation, imitation strategies and

existence of new/unfamiliar environments.

It is further evident that the early studies on confusion have been characterised by the

‘information overload’ paradigm (Jacoby et al, 1974a & b; Russo, 1974; Summers, 1974;

Wilkie, 1974; Jacoby, 1977; 1984; Scammon, 1977; Malhotra 1982; 1984a & b), a topic

that still draws much attention as ‘choice overload’, ‘assortment overload’ or

‘hyperchoice’2 (Mick et al., 2004; Scheibehenne et al., 2010). More issues like product

similarity (Loken et al., 1986; Foxman, 1990; 1992; Kapferer, 1995) and store layout and

aesthetics (Berlyne, 1960; Kotler, 1993) have long preoccupied the literature offering the

foundations to the concept of confusion.

a. Information and Product Overload

The earliest stream of consumer literature dealing with overchoice (Jacoby et al, 1974a &

b; Russo, 1974; Summers, 1974; Wilkie, 1974; Jacoby, 1977; 1984; Scammon, 1977;

Malhotra 1982; 1984a & b) has been preoccupied with public policy and has left many

questions unanswered. It could be more precisely described as a ‘dispute’ on the

conditions under which the information overload phenomenon actually exists. This stream

of research is consistent with the idea of human limited processing capacity (Bettman,

1979; Malhotra, 1982) and the theory of bounded rationality (Simon, 1955) and

challenges the idea of the consumer as the ‘homo economicus’, who according to

traditional economic theory (as in Simon, 1957), is supposed to possess faultless

information on every aspect of a purchase, can easily identify his/her preferences and does

that without any cognitive or time limitations. The idea of information overload is

2 Alvin Toffler coined the term ‘overchoice’ in his book Future Shock (1970) to describe a world in which

there was too much choice to make optimal, satisfying decisions (as in Gourville & Soman, 2005).


44

contrary based on the assumptions that consumers possess a finite capacity to encode the

choice options and the information available with these choices (Simon, 1955; Miller,

1956; Malhotra, 1982) leading to poor decision-making and dysfunctional consequences

like stress and anxiety. These dysfunctional consequences, according to Keinan (1987)

make people unable to compare and contrast many different alternatives because their

attentional capacity is depleted.

On such grounds, traditional approaches (e.g. Jacoby et al., 1974a; b; Wright, 1975;

Malhotra, 1982; Bettman et al., 1990) to measuring the amount of information provided to

consumers involve simple counts of the number of alternatives and attributes in a choice

set and define information overload by comparing the volume of information supplied

with the information processing capacity of an individual (Eppler & Mengis, 2004). For

example, the seminal study by Jacoby et al. (1974a) operationalised information overload

as ‘information quantity’ and measured consumers’ best brand choice decision based on

idiographic reports on attribute preferences. According to the study information overload

occurred when more information led to the negative consequences of decreased choice

accuracy compared to each consumer’s stated ideal. The results indicated evidence of

information overload and more importantly the study signified that consumer felt better

when provided with more information but in reality made poorer purchase decisions.

Although the potential of information overload was never really questioned, the Jacoby et

al. (1974a) data were re-analysed (Russo, 1974; Summers, 1974; Wilkie, 1974). The re-

analysis of results showed no evidence that a larger number of alternatives led to lower

choice accuracy due to information overload. Rather, the reduced choice accuracy,

originally reported, was explained as an artefact of failing to account for the higher

probability of picking the wrong brand in big rather than small set sizes, meaning that a

chance factor was implicated in the process (Wilkie, 1974). Indeed, Jacoby (1977) later


45

agreed that their study might have produced some ambiguity. However, subsequent

researchers like Malhotra (1982) demonstrated the negative effect of information load on

consumers’ choice quality and choice satisfaction.

In this spirit, it has been illustrated that the function of information overload can well

follow the inverted U-curve as in figure 3.2.

Figure 3.2 Information overload as the inverted U-curve

Source: Eppler & Mengis, 2004.

Subsequent research on the topic of information overload continued with a shifted

emphasis on quality of information provided, conceptualised as either importance of an

attribute (Keller & Staelin, 1987), information structure (Lurie, 2004) or diversity of

information dimension and repetitiveness (Hwang & Lin, 1999).

This stream of research has expanded to exploration of assortment size and type, recently

investigating aspects like how the display of an assortment influences consumer

perception of it (Broniarczyk et al., 1998), consumer satisfaction with the presentation

format and level of customisation (Huffman & Kahn, 1998) and assortment entropy

(complexity) and density (attribute variability/similarity) (Fasolo et al., 2009). The results


46

of this last study support the notion that a choice made from a large assortment is not

necessarily tyrannical. Other aspects like intense entropy and density of a specific

collection of products could however turn it to such. Hence, in several cases a simple

rearrangement of the shopping environment could mean reduction in density and entropy,

which could have better results compared to a reduction in SKU quantity.

In conclusion, although there have been lively debates regarding the topic, this stream of

research clearly shows that high cognitive load has detrimental effects on consumer

decision-making, consumption and general well-being (Schwartz et al., 2002; Carmon et

al., 2003; Baumeister & Vohs, 2003; Broniarczyk, 2008). A high cognitive load decreases

consumer choice accuracy and has negative influence on consumers’ psychological states

and energy which can lead to confusion, cognitive strain as well as lower decision

satisfaction (Mitchell & Papavassiliou, 1999; Mick et al., 2004; Broniarczyk, 2008).

b. Product Similarity

Other early papers were very much preoccupied with a conception and appreciation of

brand confusion caused by imitation strategies (Foxman et al., 1990; 1992). Aspects like

brand packaging similarity (Loken et al., 1986), own brand look a-likes and imitation

(Kapferer, 1995; Rafiq & Collins, 1996; Balabanis & Craven, 1997) and trademark

infringement problems (Miaoulis & D’ Amato, 1978), have been investigated. Such

papers adopt the position that confusion is mainly a subconscious phenomenon and in

effect stimulus generalisation.

The contribution of brand similarity confusion papers is twofold. Firstly, these studies

provide sufficient experimental evidence of consumers’ vulnerability to similarity issues

which can lead to reduced market transparency and confusion. Secondly, these studies


47

provide the basis for the investigation of several proposed antecedents of the likelihood

and level of confusion each consumer will experience.

This stream of research continues to explore new areas (Burt & Davis, 1999; Hoch et al.,

1999; Lomax et al., 1999; Warlop & Alba, 2004; Walsh & Mitchell, 2005a; Miceli &

Pieters, 2010) like consumers’ personal vulnerability to perceive similarity in the market

place (Walsh & Mitchell, 2005a) or a distinction between visual/attribute based and theme

based copying (Miceli & Pieters, 2010).

3.4.2. Deeper Investigations of Confusion

Mitchell and Papavassiliou (1999) have been the first to introduce a broader conception of

confusion as a phenomenon which extends to all four marketing Ps, laying its foundations

in three main sources which tie in with cognitive psychology: overchoice of products and

stores, similarity of products and brands and finally, ambiguous, misleading or inadequate

information. They presented confusion a primarily conscious phenomenon with

implications for consumers’ shopping behaviour (mainly in terms of the adoption of

coping strategies, Mitchell & Papavassiliou, 1997). Despite the aforementioned

introduction of this holistic view of consumer confusion, subsequent research continued to

take a product focused approach to the phenomenon, with few exceptions in recent years.

Chryssochoidis (2000) using theory spanning from product differentiation, attitudes

toward organic foods and consumer confusion discusses how the acceptance of newly

launched products can suffer due to confusion as consumers are insensitive to, tend to

overlook and not really understand the differences between the late introduced items and

existing products. The findings of the study do seem to reflect a substantial problem of

confusion for late introduced differentiated products, which is mainly depicted as slow

diffusion of such products lasting in many cases for years. Turnbull et al. (2000) focus on


48

the effect of confusion on information search behaviour in the dynamic

telecommunications market. In this context, this research identified complex structures

both at the level of external environment and sources of information which cause

ambiguities, overload and misinterpretations resulting in confusion, perceived risk and

risk reduction strategies. The researchers took into account consumers’ specific

characteristics and they also acknowledged that the degree of perceived complexity,

overload and misinterpretation might vary according to individual cognitive ability.

However, this research is, yet again, limited with respect to the scope of consumer

confusion as the main focus has been information search behaviour, consumers’ perceived

risks and once more confusion triggers in the mobile phone market (see figure 3.3 below).

Figure 3.3Conceptualisation of confusion in the mobile phone industry

Source: Turnbull et al., 2000 (p. 159)

External Environment

Changing Factors

Company

Technological Changes

Government Regulations

Market Factors

Technology

Service nature of industry

Supplier structure

Sources of information

Word of Mouth, adverts, salespeople, TV,

newspapers, magazines, consumer

groups, Health Bodies etc.

Internal Environment

Information processing

Consumer confusion

misunderstanding/

misinterpreting/of info leads to

different degrees of confusion

Perceived Risk

Volatile pricing structures, price

offers, promotion strategies

PURCHASE DECISION

NO

YES

NO


49

New insights were provided by Mitchell and Kearney (2002), who although focused on

legally acceptable measures of similarity confusion and legislation infringement issues,

they were the first to argue that consumer confusion is an attitude and consequently has an

affective, cognitive and behavioural component. This idea was taken up again later in a

conceptual paper by Mitchell et al. (2005) who presented a comprehensive review of the

confusion related literature and developed a conceptual model which depicted confusion

as an evaluative, attitude-like phenomenon, which possessed an affective, cognitive and

behavioural component. This model has not however identified any connections or

relationships among these different elements of confusion. Yet, until now there has been

no empirical research published which explores confusion as a belief or attitude.

The latest works on consumer confusion, provide definitions of the phenomenon which

are broader in the sense that they consider different potential consumer confusion triggers.

To start with, Schweizer (2004) adopted an approach based on environmental psychology

(Berlyne, 1960; Mehrabian & Russell, 1974; Donovan & Rossiter, 1982). The aim was to

conduct an empirical and conceptual analysis of all consumer confusion triggers in a retail

store environment. This approach was well justified based on theories like the Optimal

Stimulation Level (Raju, 1981) and the general principles of the S-O-R model—Stimulus-

Organism-Response—(Mehrabian & Russell, 1974). This approach identifies a direct

connection between environmental confusion and consumers’ affect. Using a mixed

method approach, the author provided firstly a good descriptive appreciation of all

environmental consumer confusion triggers along with relevant reduction strategies.

Figure 3.4 (as in Schweizer & Rudolph, 2004) summarises these triggers which evidently

correspond to retailers’ mechanisms for differentiation. Secondly, this research

operationalised confusion as a measure of ‘information rate’ and provided a confirmation


50

of the connections between negative emotional states and confusing environmental

stimuli. Negative affective conditions were measured at the aggregate level using

consumers’ emotional statements derived from a preceding focus groups study. This scale

included a diverse compilation of emotional accounts, ranging from emotional judgements

such as ‘nicht ernst genommen’ (transl.: not taken seriously), ‘überrumpelt’ (transl.:

overpowered) to primary emotions such as ‘verärgert’ (transl.: angry) and ‘gestresst’

(transl.: stressed) as in Schweizer, 2004 (p. 134). In addition, the quantitative approach

utilised the ‘shopping list approach’ (see also Titus & Everett, 1996). According to this

technique, consumers are asked to find and buy specific products from a predetermined

shopping list; such products might not be part of consumers’ daily routine. They are then

asked to answer specific questions on the experience they just had (time and situation

specific approach). Although such approaches can be said to lack ecological validity,

researchers who have utilised them argue that consumers are allowed to be confronted

with the overall complexity of a shopping occasion forcing them to breach their standard

routines.


51

Figure 3.4 Consumer confusion triggers

Source: Schweizer & Rudolph, 2004 (translated into English by the author).

A research article published as the result of this work (Schweizer et al., 2006) has focused

on the development of a multidimensional scale for consumer confusion which adopts the

same theoretical background and allows for a holistic appreciation of the retail

environment. The scale consists of six dimensions; stimuli variety (incl. similarity),

novelty, complexity, conflict, comfort and reliability. Table 3.1 describes the factors and

items of the scale.

Store Layout

Rearrangement of shelves

Unclear arrangement

Price

Unclear value for money

Missing outline of pricing actions

Assortment

Product variety

Product similarity

Assortment changes

Missing quality symbols

Location

Type of clientele

Complicated to reach

Technology

New technological applications

Complex loyalty cards

Consumer

Confusion

Market cultivation and promotions

Too much information

Diffuse labelling

Lack of trustful communication

Service

Product or information gaps on shelves

Long queues

Personnel

Unqualified personnel

Imbalanced advices


52

Table 3.1The 'environmental approach' to consumer confusion

Variety

To me, there are too many products to choose from.

It is not easy to choose between similar products.

Products often seem similar but are not.

Seven points

likert scales

(agree-disagree

format)

Novelty

Products in favour are often replaced too frequently.

Products I am looking for are often missing in store

shelves.

I often have to get used to new labels/promotion

programs.

Shelves or products are replaced very often.

Products’ packaging are often replaced.

Prices for products are often subject of frequent

changes.

Services are often subject of chance.

Complexity

I have a hard time to understand product descriptions.

I am not sure what the numerous labels are standing for.

Often I am uneasy if a more expensive product is really

better than a cheaper one.

Conflict

Sometimes I am not sure if the store wants to

manipulate me with the loyalty card program.

It is critical that my personal data are available to the

store.

I believe that the store urges me sometimes to buy

something.

Comfort

I am already irritated if I am confronted with a long

waiting line at the check-out when entering the store.

Customers shopping in stores are sometimes annoying.

Crowding in a store often disables an agreeable

shopping.

Reliability

Sometimes it is difficult to judge the quality of a

product.

I often get only vague information.

Sometimes I am not sure if I can rely on the ads

messages.

The products’ pictures sometimes do not correspond

with the real product.

Sales are often only slightly cheaper than the regular

price.

I believe that the stores frequently increase prices

without informing consumers.

Source: Schweizer et al., 2006.


53

Another approach to consumer confusion was introduced by a series of studies by Walsh

et al., 2007 and Walsh & Mitchell, 2010. The aim of the studies was to validate a scale of

consumers’ general tendency/proneness to be susceptible to stimuli like information

overload or ambiguity and provide evidence of how it could affect several consumer

behaviours, like purchase postponement, loyalty behaviour or word of mouth. They dealt

with a de-contextualised kind of confusion which as an individual trait originates in

personality research. This understanding of confusion challenges and at the same time

supplements the reasoning behind situation specific confusion, which in accordance with

the writers, is highly stimulus dependent. Three consumer confusion proneness traits were

identified as relevant for explaining confusion proneness; that is perceived similarity,

overload and ambiguity (in accordance with Mitchell & Papavassiliou, 1999 and Mitchell

et al., 2005a). Accordingly perceived similarity was described as ‘the perception that

different products in a product category are visually and functionally similar’ (Walsh et

al., 2007, p. 702). Overload confusion ‘consumers’ difficulty when confronted with more

product and market information and alternatives than they can process’ (Walsh et al.,

2007, p. 704) and ambiguity confusion as ‘consumers’ tolerance for processing unclear,

misleading or ambiguous products, information and advertisements’(Walsh et al., 2007,

705). Table 3.2 describes the factors and items of the scale.


54

Table 3.2 The ‘cognitive approach' to consumer confusion

Similarity

Due to the great similarity of many products, it is often

difficult to detect new products.

Some brands look so similar that it is uncertain whether

they are made by the same manufacturers or not.

Most brands are very similar and are therefore hard to

distinguish.

Most brands look so similar that it is difficult to detect

differences.

Five points

likert scales

(agree-disagree

format)

Overload3

I do not always know exactly which products meet my

needs best.

There are so many brands to choose from that I

sometimes feel confused.

Due to the host of stores, it is sometimes difficult to

decide where to shop.

The more I learn about grocery products the harder it

gets to choose the best.

Ambiguity

Products such as CD players or VCR often have so

many features that a comparison of different brands is

barely possible.

The information I get from advertising often is so vague

that it is hard to know what a product can actually

perform.

When buying a product I rarely feel sufficiently

informed.

When purchasing certain products, such as a computer

or hifi, I feel uncertain as to product features that are

particularly important for me. Source: Walsh & Mitchell, 2005a; Walsh et al., 2007; Walsh & Mitchell, 2010.

The results support the idea that confusion proneness is a multidimensional phenomenon

which impacts purchase postponement and loyalty (Walsh et al., 2007), word-of-mouth,

trust and ultimately consumer satisfaction (Walsh & Mitchell, 2010)– nevertheless, each

one of the three confusion proneness traits was found to have a differential impact on the

outcome variables.

3 Items are chosen to reflect the overload factor introduced by Sproles & Kendall (1986).


55

3.5. Consequences of Consumer Confusion

Consumer confusion has been characterised as a hygiene factor (in accordance with

Herzberg’s theory of motivation): its presence has been repeatedly connected to negative

consequences such as dissatisfaction; however, its absence does not motivate individuals

to purchase more nor does it necessarily lead to satisfaction (Mitchell et al., 2005).

Mitchell et al., 2005 divided the strategies consumers use in order to cope with confusion

into two categories. The first of these strategies is to simply abandon the purchase all

together and the second comprises of a number of confusion reduction strategies, like

narrowing down the alternatives or sharing the decision with a knowledgeable other.

Beyond the above coping strategies, confusion has been connected to adverse outcomes.

Even though, no study has implemented a comprehensive investigation of the outcomes of

consumer confusion (Mitchell et al., 2005), confusion has been connected with results like

decision avoidance or postponement (Tversky & Shafir, 1992; Greenleaf & Lehmann,

1995; Dhar, 1997; Mitchell & Papavassiliou, 1997; Huffman & Kahn, 1998; Anderson,

2003), status quo and omission bias (one manifestation of status quo bias could be an

increased loyalty towards existing brands) (Ritov & Baron, 1992), negative word-of-

mouth (Tumbull et al., 2000; Walsh & Mitchell, 2010), dissatisfaction (Malhotra, 1982;

Foxman et al., 1990; Walsh & Mitchell, 2010), cognitive dissonance (Mitchell &

Papavassiliou, 1999), shopping fatigue (Mitchell & Papavassiliou, 1997), increased levels

of reactance (Settle & Alreek, 1988), decreased loyalty and trust (Walsh & Mitchell,

2010) and even confusing other consumers (Foxman et al. 1990; 1992).

It is evident that the question of ‘exit, voice or loyalty’, in accordance with Hirschman’s

treatise (1970) remains and seems to find an application on such states like consumer

confusion.


56

3.6. Demarcation of the Terms

Moving to a broader topic, several psychological states which have found wide

application in psychological and consumer behaviour research seem to share theoretical

grounds with confusion. In this section, a demarcation of the terms will be attempted in

order to clarify the distinction but also the relationship between consumer confusion and

a) cognitive dissonance, b) uncertainty and perceived risk and finally, c) irritation/

frustration.

3.6.1. Cognitive Dissonance

Cognitive dissonance theory and research dominated social psychology from the 1950s

until the 1970s (Cooper, 2007). The theory is based on Leon Festinger’s suggestion that

when an individual holds two or more elements of knowledge (cognitions) that are

inconsistent with one another, a state of discomfort (dissonance) is created. Festinger

(1957) further indicates that persons are motivated by the unpleasant state of dissonance

to engage in ‘psychological work’ so as to reduce the inconsistency, and this work will

typically support the cognition most resistant to change.

In 1956, Brehm examined dissonance theory’s predictions for post-decision processing.

According to the theory, following a decision, all of the cognitions that favour the chosen

alternative are consonant with the decision, while all the cognitions that support the

rejected alternative are dissonant. The greater the number and importance of dissonant

cognitions and the lesser the number and importance of consonant cognitions, the greater

the degree of dissonance experienced by the individual. Thus, in a decision situation,

dissonance is typically greater the closer the alternatives are in attractiveness, as long as

each alternative has several distinguishing characteristics. Dissonance caused by a


57

decision can be reduced by changing attitudes or by viewing the chosen alternative as

more attractive and/or viewing the rejected alternative as less attractive.

Comparing the two states, both cognitive dissonance and confusion have a substantial

influence to the way consumers experience shopping situations due to the uncomfortable

state they might find themselves into. However, confusion acts at the point of decision by

making decisions more difficult and has been described as the cause or multiplier of

dissonance, which in consumer environments is usually a post-decision process (Mitchell

& Papavassiliou, 1997). Conclusively, dissonance is an unpleasant state which occurs

after a difficult shopping decision has already taken place rather than at the context of

purchase like confusion.

3.6.2. Buying Risk and Uncertainty

The concepts of perceived risk and uncertainty have established their own research

traditions in consumer behaviour research (Mitchell, 1999). The two terms have been used

interchangeably by consumer behaviour researchers but Knight, already in 1948, coined a

distinction between the terms. According to Knight (1948, p. 19–20), risk has a known

probability while uncertainty exists when knowledge of a precise probability is lacking.

However, known probabilities are extremely rare and consumers are very unlikely to think

in terms of them. A further demarcation of these two terms goes far beyond the scopes of

this report; it is suffice to argue that both risk and uncertainty approaches share the same

overall characteristics, pointing to: the probability of a loss, a perceived non-agreement

between expectations and reality and the subjective feeling of possible unfavourable

consequences (Urbany et al., 1989; Dowling & Staelin, 1994; Mitchell, 1998).

A linguistic-psychological analysis by Storm and Storm, (1987, p. 813) argues that the

two groups of emotional words, one characterised by confusion (confusion, baffled,


58

bewilderment, puzzled) and another distinct one characterised by uncertainty (uncertain,

hesitant, doubt, reluctant) are differentiated based on organisation, or control. Terms in

the confusion group convey lack of control in a cognitive sense; those in the uncertainty

convey lack of resolution rather than lack of control.

In terms of consumer behaviour, Schweizer (2004, p. 32) places the distinction between

risky and confusing situations in the context of products/situations involved. Risky

situations usually involve expected product attributes (e.g. the life time of a computer)

rather than actual product characteristics (e.g. the PC configuration). Thus consumers feel

the risk factor about attributes that are important but unknown, future aspects but feel

confused about contradicting, ambiguous or abundant actual information they receive on

different products.

It is true however that in order for consumers to reduce the degree of perceived risk,

uncertainty and the probability of negative consequences occurring, they will increase

their information search activity (Mitchell, 1992). Along with more information search,

consumers use other strategies to reduce risk like remaining brand loyal and postponing

decisions. These risk reduction strategies are nearly the same with the ones identified as

confusion reduction strategies (Mitchell & Papavassiliou; 1999; Walsh & Mitchell, 2007;

2010).

Confusion bears then connections and differences with risk and uncertainty. There is still

room for further research to clearly demonstrate the direction and nature of this

relationship but the evidence is for the treatment of these states as distinct, as far as these

can be differentiated based on the aforementioned characteristics.


59

3.6.3. Irritation/Frustration

In the domain of human behaviour, Roseman (1991) suggests that frustration will occur in

a situation in which a negative outcome happens, when a positive outcome is desired. That

is, the current situation is perceived as being caused by circumstances that are again

inconsistent with current motives, and result in punishment or absence of reward. This

description is very similar with the way confusion has been described elsewhere

(Ellsworth, 2003). However, an additional and main proposition of confusion is a lack of

understanding of the situation and lack of control of the ideal way to act. It seems then

that the two states share much in common but confusion is much more characterised by

intense helplessness/ indecisiveness while frustration more simply by bother or irritation.

3.7. Conclusion

Extending on the debate on whether marketing concepts should be evened with their

antecedents, their outcomes or are better perceived as a process (on this same point see

especially the example of consumer satisfaction- Yi, 1990 or that of trust- Mayer et al.,

1995), it seems that confusion has safely found its conceptualisation based on its

antecedents (or triggers). One of the scales examining the confusion construct derives

from the realm of cognitive psychology (similarity, ambiguity and overload) while the

other from environmental psychology which contains elements grounded on the concept

of atmospherics (stimuli novelty, variety, reliability etc.).

The current state of research debates the concept of confusion based broadly on the

following topics:

1) Confusion can be both a conscious and subconscious state. Most relevant research

claims that only the conscious part is easily accessible to the researchers (Mitchell et al.,

2005).


60

2) It is a state that until recently has been described as caused by individually perceived

environmental stimuli.

3) It occurs during decision-making situations (situation specific). It has also been

attributed the property of a personality trait, meaning that certain personalities have the

tendency to be/feel/report confused and some not.

4) Consumer confusion has been described as a cause of negative emotions (an important

response driver). The exact relationship however between confusion and emotions

comprises a neglected area of research (for an exemption see Schweizer, 2004).

5) According to the extant literature, consumer confusion leads to the avoidance of certain

stimuli, taking the form of specific reduction strategies (e.g. reliance on heuristics, status

quo bias, downsizing consideration set, choice deferral, reduced info search as in Kasper

et al., 2010) or other behavioural acts (ask help from a member of staff, change retail

store, complaint, negative word of mouth).

Following the exploration of consumer confusion, the next chapter will lead the reader to

a different chapter of consumer behaviour, that of the distinction between the contextual

and the intentional stance and the implications of intentional behaviourism. The

exploration of the two ways of doing research will lead to a meaningful bridging of the

two approaches resulting to the conceptual framework of this research.


61

4. THE BPM AND INTENTIONAL BEHAVIOURISM

4.1. Introduction

This chapter will shift the interest from consumer confusion and will examine the

proposals put forward by a new philosophical and research scheme, intentional

behaviourism. Commencing more generally by examining the propositions of

intentionality and behaviour analysis (the review will especially focus on the principles of

the Behavioural Perspective Model) it will reach an understanding of the ways that the

principles of both approaches can be combined meaningfully under the umbrella of

intentional behaviourism. Intentional behaviourism argues that through the combination of

otherwise incommensurable philosophical stances, the intentional and the contextual

stance, a better exploration of consumer behaviour can be achieved.

4.2. The Explanation of Consumer Behaviour

Marketing theory teach us (McKenna, 1991) that consumers’ needs and preferences

should be placed at the centre of all economic activity and should act as the starting point

of all NPD (new product developments), product launches, or quality concerns. Following

this logic, what is widely known as market orientation (Kohli & Jaworski, 1990), starts

with customers’ needs and should finish with their satisfaction. On the other end, the

consumer is expected to act in the marketplace and make all kind of choices based on

what is on offer and under all the influences that are imposed upon him/her. Apart from

choices, the consumer uses and disposes products or services and then gets into repeated

purchases, abandonment or even word-of-mouth with peers to support or argue against

such buys. The field of consumer behaviour covers then a lot of ground. It is:


62

‘the study of the ways and processes consumers follow in order to select, purchase, use

and dispose of products, services, ideas and experiences in order to satisfy needs and

desires’ (Solomon et al., 2010, p. 6).

Yet, this quest for needs’ satisfaction has been widely described as a process (e.g.

Solomon et al., 2010) where a number of factors have been identified as imposing an

influence on consumers. Kotler & Armstrong (2010) classify these influences as

following:

Table 4.1 Factors affecting consumer behaviour

Psychological (motivation, perception, learning, beliefs and attitudes)

Personal (age and life-cycle stage, occupation, economic circumstances,

lifestyle, personality and self concept)

Social (reference groups, family, roles and status)

Cultural (culture, subculture, social class system).

Source: Kotler & Armstrong (2010), p. 161-175.

A closer examination of these categories of influences reveals two main sources of

consumer behaviour ascription; consumer behaviour can be either attributed to some

characteristics of the individual (cognitive perspective) or to the external socio-cultural

environment (socio-cognitive perspective). Intrapersonal characteristics (psychological

processes within the person like perception or attitudes) along with socio-demographics

form the first category, while social and cultural factors form the external environmental

influences. These groups of influences characterise the way consumer behaviour is

conceived and presented in most renowned consumer behaviour and marketing textbooks

(e.g. Solomon et al. 2010; Kotler & Armstrong; 2010). This conception of consumer

behaviour is in accordance with the principles of cognitive psychology. Cognitive

psychology, the sub-discipline of psychology exploring internal mental processes (APA,

2013), has gained an unambiguous status among consumer behaviour researchers (as in

Holbrook & Hirschman, 1982; Foxall, 1996). Based on the long tradition of cognitive


63

revolution in psychology and consumer behaviour (Neisser, 1967; 1976; Bettman, 1979)

the word ‘cognition’ has been most often defined to include such aspects as perception,

attention, implicit and explicit memory, language processing and mental processes,

networks and schemata that connect all these functions and result to problem solving

(Bettman, 1979; Plutchik, 1985; Sternberg, 1999). So immense is its influence that its

application and status goes unnoticed by researchers applying its principles, who use it as

though this is the only available stream of thinking.

In order to battle this prevalence, some researchers have proposed a substitute to cognitive

psychology grounded on an interpretive approach to consumer behaviour and based on the

examination of the ‘experiential’ aspects of consumption (e.g. Holbrook & Hirschman,

1982). Still, this stance does not act at an ultimately different level from cognitive

psychology. The accounts provided by interpretive approaches still centre around the

individual consumer but differ from cognitive psychology mainly on three aspects: 1)

their explanation extends to more individual characteristics, other than memory,

perception and attitudes, like fantasies and emotions, 2) the research endeavours expand to

more than the usual everyday products or situations, namely to ‘extraordinary

experiences’ (e.g. Arnould & Price, 1993) and 3) more freedom is provided to the

researcher to take a more active role during the research process (e.g. Hirschman, 1986;

Thomson et al., 1990). Although this stream of research endorses the power of situational

influences (maybe more accurately cultural influences) on consumers, only an explanation

of consumer behaviour based on behaviourism or behaviour analysis achieves to shift the

interest from intrapersonal processes to the situational influences upon choice and

consumption (Foxall, 1986; 1987; 1988). This approach extends the theoretical

understanding to the ways the environment shapes consumer behaviour over time. In

essence the cognitive and behavioural approaches describe different conceptualisations of


64

learning. The focus to the fundamental process of consumer learning (Skinner, 1950) can

illuminate the main differences between the two approaches.

4.3. Ways of Learning (Cognitive versus Behavioural)

The way consumers store information in memory (i.e. what they know, think and feel

about products, brands or situations) is through the process of learning. Most of the

changes in behaviour are a result of experience and most experiences result in learning.

Thus, the information acquired through learning forms the basis of consumer behaviour

(Catania, 1998; Foxall et al., 1998; Solomon et al., 2010). This is why the definition of

this process is of the outmost importance in the examination of human nature and can

assist in the process of discriminating between cognitive and behavioural approaches of

understanding.

Learning possesses then a central role in both the cognitive and behavioural traditions and

Foxall et al., (2008) distinguish between the cognitive and behavioural ways of explaining

learning. Cognitive learning theories view this process as a ‘conscious mental activity’

which is very much in accordance with the predominant view of consumer behaviour as

information processing. Here the focus is on the way information gets from the external

environment into long term memory, usually causing internal, mental changes (in the form

of beliefs, attitudes or schemata) and ultimately influencing consumers’ decision-making.

This process is determined by the way that a buyer can function intellectually. In its

simplest form consumers are claimed to learn by repetitive exposure to stimuli which

means that they simply memorise without paying much attention (low attention

processing). Due to repetitive exposure consumers may develop weak beliefs for brands,

firms or logos. One of the most significant applications of this principle is in advertising.

This kind of learning indicates that consumers do not need to cognize or process


65

advertisements; simple repetition can create a memory trace which can influence

consumer behaviour. Research in advertisement which is based on the mere exposure

effect (Zajonc, 1968), that is peoples’ tendency to develop a preference for things they are

frequently exposed to, has explored the effect of mere exposure to consumers’

judgements, attitudes, preferences and behaviour (Grimes & Kitchen, 2007).

Other forms of cognitive learning might take the social learning perspective, where

consumers watch others behave and then apply these principles in their own life (Bandura,

1977 cited in Foxall et al., 2008). Finally, the most influential of all the cognitive

approaches, the information processing concept is also the most comprehensive approach

and includes elements of both the two aforementioned simpler forms of cognitive

learning. This kind of learning requires extensive mental effort and time from the

consumer and encompasses a wide range of cognitive activities. Many elaborate models

of information processing have been developed, especially in advertisement research

(Petty et al., 1983; Greenwald & Leavitt, 1984 cited in Foxall et al., 1998, p. 79; Maclnnis

& Jaworski, 1989). Such models usually present learning as an elaborate task comprising

of elements like exposure, attention, comprehension, several cognitive and more recently

emotional responses and finally, storing of information in memory. This elaborate chain

of tasks usually leads to attitude or beliefs formation or change.

Behavioural approaches at the other end, describe learning as ‘largely unconscious

changes in overt and verbal behaviours’ as a result of consumers’ experiences and

environmental influences (Foxall et al., 1998, p. 76; p. 90). Consumers learn to form more

sophisticated beliefs through the largely unconscious processes of classical and operant

conditioning. Conditioning occurs through interaction with the environment, while

environmental influences are considered the sole forces responsible for shaping our

behaviour (Skinner, 1972 cited in Markin & Narayana, 1975).


66

A more detailed presentation of the two approaches will be achieved in the remaining of

this chapter.

4.4. Behaviour Analysis

Behaviour analysis is the field of philosophy, research and application that encompasses

the experimental analysis of behaviour (experimental research designed to add to the

knowledge about behaviour), applied behaviour analysis (focusing on applying these

behaviour principles to real world situations), operant psychology, operant conditioning,

behaviourism and Skinnerian psychology (Vaughan, 1989 cited in Foxall, 1996; Foxall,

2001). In marketing-oriented economies, consumer behaviour analysis is principally

concerned with ‘human behaviour in naturally occurring settings that are subject to

marketing influences’ (Foxall, 2002 as in Foxall, 2013, p. 105). It is based on the

principles of behaviourism which as a distinct field of psychology and focuses on

behaviour as a main subject matter and has been developed based mainly on the work of

three influential psychologists, Ivan Pavlov, John Watson and Burrhus Frederic Skinner.

Pavlov’s work (1927 cited in Macklin, 1986) is the basic paradigm introducing and

advocating classical conditioning. The theory behind classical conditioning has been

called the stimulus substitution theory (Mazur, 2006, p. 65). Pavlov’s experiments are

famous and so are the dogs that were the subjects of the experimental conditions. What he

did is to arbitrarily select a stimulus, a bell or a metronome, to be presented to a hungry

dog. The metronome acted as the Conditioned Stimulus (CS), which was then followed by

food, the Unconditioned Stimulus (US). The US caused the dog to salivate, the

Unconditioned Response (UR). With pairing the metronome (CS) with the food (US), the

dog responded to the metronome (CS) by salivating, now called the Conditioned

Response (CR), even in the absence of food. When the CS is presented before the US, this


67

is called forward or traditional conditioning, when these stimuli are presented

simultaneously this is simultaneous conditioning and when the US precedes CS this is

known as backward conditioning (Macklin, 1986). This psychology of ‘conditioned

reflexes’ (as in Foxall et al., 2008) gained in popularity during the first decades of the 20th

century, especially in the USA, where Watson reacted against the extended use of

introspective methods among psychologists. In his 1913 article ‘Psychology as the

behaviourist sees it’ he extensively argues that psychology must discard all reference to

consciousness and that behaviour should be the objective point of reference of all

psychological measurements (Watson, 1913).

The principles of classical conditioning have been widely applied to marketing, especially

to advertisement research. This is usually done through experimental approaches where

advertisement features like music, source or content are paired with products or brands to

examine if classical conditioning has an effect on desirable advertisement results.

However, no matter the fact that the basic value of behaviourism is the exploration of

behaviour, most of the studies utilising the classical conditioning approach in

advertisement research, usually use attitudes or preferences as their dependent variables

(as in DiClemente & Hantula, 2003). The same authors (DiClemente & Hantula, 2003),

following a detailed literature review on the use of the behavioural perspective in

consumer research argue that classical conditioning procedures have been found to have

an effect on attitudes and other similar indirect measures of behaviour. However,

establishing the effect of classical conditioning on actual choice has proved difficult. One

of the initial experiments on the effects of classical conditioning on choice used photos of

blue and beige pens (Conditioned Stimulus) and connected these to either liked or disliked

music (Unconditioned Stimulus) (Gorn, 1982). At the end when consumers were given the

chance to choose the blue or beige pen most participants chose the pen connected to the


68

liked music. This result indicates that they did not base their immediate choice on colour

preference or their personal characteristics but chose based on the connection with the

unconditioned stimulus. The evidence, connecting classical conditioning with choice,

remains however, sparse.

To finish with, Skinner introduced the concept of operant conditioning, in which

reinforcement (or punishment) leads to a desired behaviour. The main difference of

operant with classical conditioning lies on the fact that much everyday behaviour is not

educed by a specific stimulus as proposed by classical conditioning (Blackman, 1974, p.

49; Mazur, 2006, p.118). Rather, in the presence of stimuli, creatures, especially humans

might choose to act or defer acting. It seems then that engaging or not in ‘voluntary’

behaviours comes in contrast with the ‘involuntary’ behaviours that are part of

unconditioned and conditioned reflexes. In simple terms, operant conditioning is a kind of

learning in which an individual’s behaviour is shaped by its consequences. The

behavioural consequences are shaped by the influence of reinforcers or punishers, which

are the core tools of operant conditioning; due to these consequences behaviour can

change in shape, strength or frequency. The general aim of the research in operant

conditioning is then characterised by the attempt to portray general principles and

functional relationships that could predict the kind of non-reflexive behaviour produced

and to reveal the situational conditions under which such non-reflexive behaviour can be

shaped (Mazur, 2006, p. 119). The principles of operant conditioning have resulted in the

development of one theoretical stance of research the ‘contextual stance’ (Foxall, 1999b;

2000) and one influential theory of consumer choice and behaviour, the Behavioural

Perspective Model- BPM (Foxall, 1987; 1990). The principles of operant conditioning and

the BPM of choice and consumption will be analysed in more detail in the following

sections.


69

4.5. Operant Conditioning

Nord and Peter (1980) were among the first to theoretically introduce the principles of

operant conditioning to marketing research. Contrary, classical conditioning has found

wider applications to consumer research. This is true because the classical conditioning

approach has shorter operational time when compared with operant conditioning. In

addition, classical conditioning is much more easily integrated into the cognitive stream of

research with variables like attitudes and preferences widely used (DiClemente &

Hantula, 2003).

Foxall in 1987 emphasised the importance of operant conditioning with the introduction

of the principles of radical behaviourism (the formal designation of Skinner’s philosophy

of science) in consumer behaviour. The same author (Foxall, 2001) describes that the role

of behaviour analysis in marketing and consumer research is to explain behaviour in terms

of the consequences it produces but also the rewards and punishments that are contingent

upon it. Economic behaviour is portrayed as operant because it operates in the

environment to produce consequences (Foxall, 1992b, p. 385). The rate and form of

behaviour depends on the consequences such behaviours had produced in the past, mainly

in terms of reinforcers, consequences that appear to strengthen a behaviour or punishment,

consequences that have aversive effects. There are however some consequences which are

neutral, meaning that these do not have an effect on subsequent behaviour. Antecedent

variables are also implicated in this process. These are the stimuli that signal the

reinforcement (or punishment) contingent upon the specific setting and thus in their

presence the consumer discriminates, by performing those behaviours that have been in

the past reinforced and avoiding those that have been punished. So, in its simplest form

the ‘three term contingency’ initially introduced by Skinner implicate the discriminative


70

stimuli (antecedents), the behaviour and the consequences in the form of reinforcers or

punishment. This simple sequence of behaviour is depicted in figure 4.1.

Figure 4.1 The A, B, C of causation: antecedents (discriminative stimuli),

behaviour and consequences

Source: Foxall, 1992b, p. 386.

Figure 4.1 is a way to simply depict the central explanatory device of radical

behaviourism as introduced by Skinner in 1953 (cited in Foxall, 1996a). The three terms

contingency:

SD

R St

and

S

D R S

p

where SD

is a discriminative stimulus, R is a response and St/p

is a reinforcing or punishing

consequent stimulus. In more detail this paradigm explains behaviour in terms of the

operant response- any arbitrarily defined bit of behaviour, a discriminative stimulus-

which signals the likelihood of reinforcement and punishment of behaviour that is

contingent upon the performance of a particular behaviour (Michael, 1980), the

A : B : C

A: Discriminative Stimuli (Antecedents)

B: Behaviour

C: Consequences (reinforcers, punishers)


71

consequences of an operant response- which can be either reinforcing or punishing and

the individual’s learning history- which is the sum of a consumers emitted behaviour and

summarises the cumulative contingencies of reinforcement and punishment under which

an individual has behaved in the past.

4.6. Kinds of Reinforcement (and Punishment)

Due to the undisputed value of reinforcement and punishment as one of the building

blocks of operant conditioning, several classifications have been proposed.

4.6.1. Primary and Secondary Reinforcers

The usual distinction of reinforcers is between primary and secondary (Blackman, 1974,

p. 93; Foxall, 1997a, p. 85). Primary reinforcers are effective from birth and usually apply

to all species; food and water are two examples. The characteristic of primary reinforcers

is that their influence on behaviour does not depend on other reinforcers. They act

naturally to determine the rate of behaviour. On the contrary, the power of secondary

reinforcers is acquired through an individual’s experience. Their influence on the rate of

behaviour depends upon their pairing with primary reinforcers. Money is the most

common example of secondary reinforcers, as it is used to obtain many primary

reinforcers.

4.6.2. Contingency-Derived and Rule-Derived Reinforcers

Another useful distinction is the one between contingency-derived and rule-derived

reinforcers (Foxall, 1997a). In cases that rules act as reinforcers, contingency-derived

reinforcers can be both primary and secondary in nature and are apparent in the

contingency shaping of behaviour (contingency-governed behaviour- Skinner, 1969).

These are generally associated with pleasurable events and this is the reason why such

reinforcers are usually utilitarian/ hedonic in nature. On these same grounds, punishment


72

can derive from events which are unpleasant. It should be noted though that the

assignment of pleasure as a defining characteristic of reinforcers is not as clear as

described here and has been the subject of long debates. The verbal (other or self) rules at

the other end are not defined by nature, thus are usually classified as secondary reinforcers

and their effect on behaviour is mediated by others (but as explained below ‘others’ might

be the individual him/herself). They do have an informational value and their level of

influence depends on something else—the type of rule, the context in which the rule is

provided, the level of trust towards the person which creates or dictates the rule and the

level of experience with the specific rule of the person who behaves (Peláez & Moreno,

1999). In the same sense, punishment would be the result of no compliance with the rules

(or self-rules as such).

The interest in the distinction between contingency and rule-governed behaviour can be

traced in Skinner (1966) who argued that in humans, who are verbal creatures,

reinforcement and consequently behaviour could arise: 1) from the direct contact with

environmental contingencies (contingency-shaped) or 2) from verbal descriptions of these

contingencies provided by the individual or others, which he termed rules.

a. Rule-Governed Behaviour

Verbal rules, which result in rule-governed or instructed behaviour (RGB) (Skinner, 1966

cited in Törneke et al., 2008; Skinner, 1969; Catania, 1986; Catania et al., 1990; Foxall;

1997a; Törneke et al., 2008), have been introduced, beyond their reinforcing power, in an

attempt to explain complex human behaviour which does not always follow the three term

contingencies. Rules can act as instructions and are effective as long as they are either

specified in rules (in essence social rules or norms), or result from the verbal activity of a

speaker or from rules or self-rules to which an organism has adhered throughout its

history (Foxall, 1997a). For example, society rules that money has a dual function: it gives


73

people the power to exchange and it can be further perceived as a measure of prestige and

success. Similarly, upon the instruction: ‘Put your coat on and you will be warm’ a child

will put on his coat without further contingencies (Törneke et al., 2008).

To date there is a debate over the actual role of those rules in behaviour analysis because

such rules may enter any kind of behavioural relationship (Cerutti, 1989). The most

prominent functions suggested for the rules in question are those of reinforcers (Foxall,

1997a), those of verbal discriminative stimuli that can take the place of the contingencies

themselves and strengthen or weaken behaviour (Skinner, 1969; Galizio, 1979; Baum,

1995; Okouchi, 1999) or those of function-altering contingency specifying stimuli, which

alter the function of other stimuli in a manner analogous to operant conditioning (Blakely

& Schlinger, 1987; Schlinger & Blakely, 1987; Schlinger, 1993). Following the

instruction: ‘When you hear the bell, leave the room’, behaviour will only be performed as

a result of the sound of the bell and not as a result of the verbal rule, the bell now takes the

role of the discriminative stimulus. Subsequently, the verbal rule is having an altering

function, to turn the bell from an irrelevant stimulus to a cause of behaviour (Blakely &

Schlinger, 1987).

This multi-functional nature explains why several authors have proposed that the

terminology used to describe rules should always reflect the specificity of the

phenomenon which is of interest each time (Brownstein &Shull, 1985; Michael, 1986). In

a similar vein, Catania (1986) proposes that a rule should be judged and better defined

based on the level of effect it has on behaviour rather than on any other basis.

However, the disparity between those conceptualisations goes beyond simple terminology

and has implications for both theory and research practice. The conception of rules as

discriminative stimuli for example implies that their function is to evoke the behaviour in


74

question. Contrary, the latter description of rules as either reinforcers/ punishers or

function-altering contingencies implies a functional relation between a stimulus and

behaviour and opens new approaches to theory and research, as the question of how a rule

can alter/ influence the relationship of stimulus and behaviour remains to be answered.

b. Functional Units of Rules (and Self-Rules)

Another central point in the debate of rule-governed behaviour is the development of

different functional units of rules for the speaker, the listener and also the formulation of

self-based rules (Zettle & Hayes, 1982). Although speaker units of rule-governed

behaviour in the form of mands and tacts have been proposed by Skinner (1957), Foxall

(2010a, p. 82) describes that consumer behaviour researchers should be mainly concerned

with the verbal behaviour of the listener (Zettle & Hayes, 1982; Schlinger, 2008),

including cases that the listener is the same with the speaker. In this second instance, the

rules are actually self-based rules and have been described as being of importance (Foxall,

1997a; Kunkel, 1997; Foxall, 2010a) because such self-rules can be formulated in order to

guide habitual, everyday behaviour.

On these grounds, three categories of listener rule-based behaviour have been proposed

and analysed by the literature (Zettle & Hayes, 1982; Poppen, 1989). This account of

listener-based units of rule-governed behaviour is composed of pliance, tracking and

augmenting.

Rule-following that is socially mediated is known as pliance. In this case, the listener’s

behaviour is mediated by the rules of another individual, the speaker, who has the power

to reward or punish subsequent behaviour based on conformity or disobedience to the rule

(Zettle & Hayes, 1982; Foxall, 2010a). Foxall (2010a) argues that a great deal of

consumer behaviour is actually pliance. Pliance can be found in cases when someone is


75

doing what somebody else is saying in order to either comply with this person’s rules as in

a thief saying, ‘Your wallet or your life’ (Zettle & Hayes, 1982), or to obtain another

person’s favour by following his rules (Törneke et al., 2008) or in order to comply with a

rule that clearly states the reinforcing consequences of doing so (Foxall, 2010a). A child

conforms to spending his/her pocket money as instructed by a parent, following the rule

‘spending wisely, you can save more at the end of the week’. Such rules are known as

plys.

Another common category is the rule-governed behaviour which arises from rules

specified by another person, who is however not in a position to reinforce or punish

others’ behaviour (Foxall, 2010a). This time the behaviour is known as tracking, the rules

are known as tracks and it is usually the physical environment that mediates the rule

following of such rules. For example, when a passerby instructs a person the way to a

store, the speaker is in no position to supply reinforcement or punishment for getting or

not there. Success or failure to find the store depends upon progress in getting there and

reinforcement is provided by finding the store, while punishment by failing to find the

store (Foxall, 2010a). According to Glen, (1987) tracks can function as antecedents to

behaviour and are expected to have a behavioural effect (Catania, 1989).

The third unit of rule-governed behaviour is termed augmenting (Zettle & Hayes, 1982). It

is rule-governed behaviour which does not specify contingencies or consequences but

rather states emphatically the reinforcing or punishing value of the consequences specified

in the rule (Törneke et al., 2008). The rule itself has been termed an augmental. It is

possibly the most difficult and advanced type of rule-governed behaviour and it is usually

found in mixed form with either pliance or tracking (Zettle & Hayes, 1982). The results of

augmentals are mainly evident when interacting with pliance or tracking and people act on

an augmental usually where the consequences might be obvious at a subsequent time.


76

In addition to the rules which might govern listeners’ behaviour and are introduced by

others, self-instructions or self-rules appear as a special kind of rule-governed behaviour

(Zettle & Hayes, 1982; Vaughan, 1985; Zettle, 1990; Kunkel, 1997). In this instance the

speaker and the listener are the same individual. The main reason for the development of

such self-rules has been described as ‘being personal’ in the sense that an individual can

react more effectively now or in a future occasion than when based on the contingencies

alone (Zettle & Hayes, 1982). Such learned behaviour may evoke appropriate actions in

the future faster than the actual contingencies it describes (Vaughan, 1985). The rules the

person formulates act then as a learning history (history of reinforcement or punishment)

which the individual can rely on.

Another characteristic of self rules is that due to the unclear distinction between the rule

giver and follower it is more difficult to distinguish among different kinds of functional

units. Self-pliance and self-tracking range along a continuum rather than been two distinct

categories (Zettle & Hayes, 1982). In analogy to listener’s units of rule based behaviour,

self-tracking occurs when the rule is to be followed because this is a description of the

state of affairs and self-pliance occurs when the rule is to be followed simply because it

was formulated (Zettle & Hayes, 1982, p. 90).

A detailed analysis of all units of rule-governed behaviour (speaker-, listener- and self-

mediated) based on Skinner (1957) and Zettle & Hayes (1982) categorisation is provided

in Appendix 2.

c. The Importance of Rule-Governed Behaviour

The importance of rule-governed behaviour lies on the rules’ power to specify the setting-

response-outcome contingencies and ultimately consumer behaviour. In terms of the ways

the rules influence market behaviour an example is set from the way decision-making is


77

described when seen from the behavioural approach. In case an individual lacks a relevant

learning history, therefore does not have the experience/knowledge to perform a

behaviour, in cognitive terms this consumer is said to resort to ‘systematic processing’ or

the ‘central route to persuasion’ (as in Foxall, 1997a; Foxall, 2000). In an interpretation

based on radical behaviourism however, the same behaviour is said to be governed by

‘other rules’, which can be conceptualised as following social norms or social pressures in

order to act. As experience of the situation increases such ‘other rules’ are replaced by

‘self-rules’, which in cognitive terms correspond to the ‘peripheral route to persuasion’.

Table 4.2 depicts the proposed role of ‘other’ and ‘self’ rules in connection with the

experience of a situation.

Table 4.2 Behavioural and cognitive approaches to decision-making

Low experience High experience

BPM

Other rules. Consumers

lack a relevant learning

history prompt search for

other rules.

Self rules. Acquisition of a

learning history, from which self

rules can be extracted.

Elaboration

Likelihood Model Central route Peripheral route

Mode Deliberation Spontaneity

Heuristic-systematic

processing Systematic processing Heuristic processing

Source: Foxall, 1997a; Foxall, 2000.

According to Foxall (1997a; 2000) in the markets and situations which are signalled by

higher levels of experience consumers develop self-rules in the form of tracks (Zettle &

Hayes, 1982). These situations can then be described as habitual/ every-day settings and

this fact results to easier and faster decision-making. When markets have reached such a

state then the actual contingencies of the market play a more important role in decision-

making than other rule-based behaviour.

More recently, Foxall (2013) has advocated the inclusion of such terms in the expansion

of the BPM to intentional terms. Rules as will be explained are having a distinctive


78

position to play in such an exploration. As rules can easily summarise what consumers

have done in the past are good candidates to take the position of learning history and to

define the consumer situation (Foxall, 2013; Oliveira-Castro, 2013). These principles will

be further explored and applied in the exploration of this conceptual framework (see

chapter 6).

4.6.3. Utilitarian and Informational Reinforcers

Following the elaborate description of rule-governed behaviour, one further distinction

between utilitarian and informational reinforcement has been introduced (Foxall, 1992a).

Utilitarian reinforcement consists of the functional benefits and material satisfaction of

situations. These can be translated as the practical benefits of purchase and consumption

which are not mediated by other people (Foxall, 2004). It is mediated by the product or

service itself (Foxall et al., 2006). Informational reinforcement can be conceptualised as

performance feedback, an indication for the person who behaves of how well that

individual is doing. It is usually mediated by the responsive actions of others and thus

confers social status, self-satisfaction or simply denotes progress to date (Foxall, 1997a;

Foxall et al., 2006). These two kinds of reinforcers have been conceptualised as the

consequences of consumer behaviour by the theoretical model of the BPM and will be

further analysed below.

Emphasises should be placed on the fact that utilitarian reinforcement is not always

primary and informational secondary as implied until now (Foxall, 1997a). Primary

reinforcement might emphasise the utilitarian aspects of consumption but many times in

humans, there can be an informational component and secondary reinforcement can be

both utilitarian and informational. At the same time, some reinforcers, like money, can

have both utilitarian and informational value. Consequently, their interpretation depends

upon the situation examined.


79

Table 4.3 categorises the kinds of reinforcements and clarifies their distinctions and

relationships further.

Table 4.3 Sources of reinforcement

Contingency-derived Primary reinforcement Utilitarian (plus informational)

Contingency-derived

(may be rule-assisted) Secondary reinforcement Utilitarian and informational

Rule-derived Social/verbal

reinforcement Informational (plus utilitarian)

Source: Foxall, 1997a, p. 86

4.7. The Contextual Stance

In order to extend the applications of the principles of operant conditioning, Foxall

(1999b; 2000) introduced the contextual stance as a philosophical framework which can

help to accommodate the principles of the environmental determination of behaviour. The

contextual stance maintains that ‘behaviour is predictable insofar as it is assumed to be

environmentally determined’; specifically, behaviour can be predicted insofar as it is

under the control of a learning history that represents the reinforcing and punishing

consequences of similar behaviour previously enacted in settings analogous to that

currently encountered (Foxall, 1999b; 2000).

The contextual stance is one of the main theoretical propositions that the BPM (the

Behavioural Perspective Model) is based upon, which (along with the Behavioural

Economy of Consumption model- Rajala & Hantula, 2001) is an influential consumer

behaviour model, developed within the principles of behaviour analysis/ behavioural

psychology. Specifically, consumers’ learning history acts as the point where particular

settings turn into discriminative stimuli, based on the way that behaviour has been

reinforced or punished in similar settings in the past. Behaviour then becomes

environmentally determined in similar settings based on the rate of reinforcement or

punishment.


80

4.8. The BPM (The Behavioural Perspective Model)

Consumer behaviour settings can be diverse; they include retail stores of many different

types, sports events, opera performances, libraries, banks, museums, airplane journeys,

and paying insurance premiums by credit transfer (Foxall et al. 1998). In fact, it is very

difficult to think of a part of modern life that does not involve aspects of pre- during or

post- consumption activities and which does not therefore place the consumer in some

behaviour setting or another (Foxall et al., 1998, p. 205).

The analysis prompted by the Behavioural Perspective Model (Foxall, 1990)

systematically relates known patterns of purchase and consumption to the situations in

which they occur. The conceptual basis of the model is neo-Skinnerian (it has been

developed from Skinner’s operant conditioning principles) and is further based on the

doctrine of the contextual stance (Foxall, 1999). As described previously in detail, the

contextual stance argues that consumers’ learning history adds meaning to the elements

that make up the behaviour setting, transforming these settings into discriminate stimuli

that signal specific outcomes to consumer behaviours enacted previously in these

situations. It is then the intersection of learning history (based on consumer experience)

with the current setting that defines a consumer situation (Foxall et al., 1998). Although

the BPM is built on the fundamental explanatory principles of behaviour analysis, the

model has two distinct points of emphasis which make its explanatory stance different

from the original versions of radical behaviourism. The two emphasising points are (as in

Foxall, 1998): firstly, the extent to which behaviour can be manipulated varying upon the

scope of the setting within which the behaviours occur; and secondly, that in consumer

behaviour settings there are bifurcations of reinforcement which are utilitarian and

informational in nature and which have separated effects on behaviour.


81

According to the behavioural perspective model (BPM) which is depicted in figure 4.2,

aspects of consumer behaviour are then predictable from two dimensions of situational

influence:

(1) the consumer behaviour setting; and

(2) the utilitarian and informational reinforcement signalled by the setting as informed by

the consumer's learning history.

Figure 4.2 The Behavioural Perspective Model of Consumer Choice

Source: Foxall, 1996, p. 26

Namely, the consequences of consumer behaviour that stem from a consumer situation are

of three kinds, utilitarian, informational and aversive consequences which reduce the

probability of future repetition (costs or utilitarian and informational punishment).

Utilitarian reinforcement is defined as the functional benefits of consumption, while

informational are the symbolic benefits like social status, self-esteem, pride and honour.

Informational reinforcement can also be described as feedback on the level of

performance of the consumer (Foxall, 1992b; Foxall & Soriano, 2005). Thus, utilitarian

reinforcements are the direct, functional benefits of being in a situation per se but it can

also derive from owning and using products and services, while informational is an

LEARNING HISTORY

CONSUMER BEHAVIOUR-SETTING SCOPE

BEHAVIOUR

UTILITARIAN/HEDONIC REINFORCEMENT

INFORMATIONAL REINFORCEMENT

AVERSIVE CONSEQUENCES


82

outcome of socially and physically constructed aspects of the environment. According to

the BPM, informational reinforcement does not derive from the typical explanation of the

word ‘information’, but it refers to the feedback-information an individual receives on its

level of its performance (Foxall, 2010a). The root of informational reinforcement lies in

the notion of secondary reinforcement through status (see above on the explanation of

reinforcements and Foxall, 2007c). Physical stimuli that can excite the senses like exciting

packaging or colours or situations that can enhance self-esteem like driving or owning an

expensive car, are perceived as typical examples of informational reinforcement.

Based on these general principles the model identifies three interactive levels of

interpretive analysis, namely, the operant class, the contingency category and the

consumer situation (Foxall, 2010a).

4.8.1. Level I: The Operant Class

Within this context, consumer behaviour takes one of four broad forms depending on the

pattern of reinforcement, informational or utilitarian, on which it is maintained. The four

operant equifinality classes proposed are: maintenance, accumulation, pleasure

(hedonism) or accomplishment. Equifinality in this context means that all members of a

class are expected to produce similar patterns of consequences (Foxall, 2010a).

Table 4.4 Operant classification of consumer behaviour

High Utilitarian

reinforcement

Low Utilitarian

reinforcement

High Informational

reinforcement ACCOMPLISHMENT ACCUMULATION

Low Informational

reinforcement HEDONISM MAINTENANCE

Source: Foxall, et al. 1998; Foxall, 1992a and b.

Following this logic and starting with the basic of the four forms of consumer behaviour,

maintenance is routine behaviour necessary to stay alive and well, such as eating,


83

shopping for groceries or buying a newspaper (Foxall & Soriano, 2005). It also includes

the purchasing and consuming of goods that are necessary to function as a member of a

society: paying taxes, for example or even waiting at an airport terminal for a flight to

leave. All these situations are maintained by low utilitarian and informational

reinforcement. Accumulation is the planned acquisition of a series of reinforcers which

have limited utilitarian content, but which are principally informational. Such behaviour is

sustained and strengthened by the provision of further reward—interest on a bank

account, prizes exchanged for coupons, collecting loyalty points from purchasing (Foxall

& Soriano, 2005) are examples of such behaviour. Hedonism is behaviour usually

reinforced by entertainment. It is maintained by a high level of hedonic (utilitarian)

reward and a lower level of informational reinforcement. Being at a party, reading a

novel, watching entertaining videos and TV programmes or being at a cinema theatre are

examples of this situation used in past research (Foxall & Soriano, 2005). Finally,

accomplishment is defined as personal achievement and is maintained by relatively high

levels of both utilitarian and informational consequence: conspicuous consumption, which

indicates both social and economic achievement, is just an instance of this behaviour.

4.8.2. Level II: The Contingency Category

Consumer situations are further categorised based on the behaviour setting scope. The

behaviour setting scope is defined by how closed or open an environment is considered to

be. This idea of the consumer behaviour setting is derived from the work of Schwartz and

Lacey in 1988 (as in Foxall, 1999a), who studied experiments with animals. They describe

how all experiments are essentially conducted in closed settings as the experimenter is the

main controller of the conditions and reinforcers provided. However, as consumer

situations anyway tend to be much more open than any experimental context, the

consumer behaviour settings differ from one another in terms of two dimensions: the


84

locus of control and the prescribed behaviour programme (Foxall, 1999a). On these

grounds, the theory describes that all settings form a continuum from closed to open. The

relative openness or closeness of settings depends on three aspects (as in Foxall, 1999a):

Whether there are readily available alternatives to being in the specific setting.

Whether the consumer or someone else controls access to or deprivation of the

reinforcers.

Whether the contingencies are imposed by agents who are themselves not subject to

them.

Hence, in an open setting consumers feel they have discretion over which choices and

alternatives are available, while in closed settings such choices are minimised and actually

determined by other agents (outside the consumer) who are not themselves subject to

these contingencies. In most affluent and consumer-oriented economies, open settings are

common place and competition allows for an abundance of choices and alternatives

(Foxall, 1999a). For example, retail settings are so designed to allow for exploration and

choice while banks are usually designed in such a way as to minimise exploration and

fully indicate consumers how to act or where to stand. Other examples of closed

environments include such cases as airplane flights, seminars or being at a cinema.

The four operant classes which are distinguished based on the level of utilitarian and

informational reinforcement along with the level of closeness or openness of the

environments (consumer behaviour setting scope) give the eightfold categorisation of the

contingencies that according to the BP model control human behaviour. A contingency

category is a way to summarise the contingencies of reinforcement pertaining to a set of

consumer situations. The complete BPM is presented in table 4.5.


85

Table 4.5 The BPM Contingency Matrix

BEHAVIOUR SETTING SCOPE

Closed Open

ACCOMPLISHMENT

(high utilitarian, high

informational)

Contingency

Category 2

Fulfilment

Contingency

Category 1

Status Consumption

HEDONISM

(high utilitarian, low

informational)

Contingency

Category 4

Inescapable

Entertainment/Pleasure

Contingency

Category 3

Popular Entertainment

ACCUMULATION

(low utilitarian, high

informational)

Contingency

Category 6

Token-Based Consumption

Contingency

Category 5

Saving and Collecting

MAINTENANCE

(low utilitarian, low

informational)

Contingency

Category 8

Mandatory Consumption

Contingency

Category 7

Routine Purchasing

Source: Foxall, 1992a and b; Foxall et al., 1998, p. 210.

Starting again with the most basic of the contingency categories, maintenance in an open

setting refers to routine purchasing and consumption (CC7), such as habitual purchasing

of grocery items. Maintenance in a closed setting can be described as mandatory

consumption (CC8), which includes all forms of behaviour necessary to remain a citizen

(such as tax payments). Mandatory consumption has also been described in terms of the

unavoidable ‘hassles’ of everyday consumer situations, like being delayed in a long queue

at a bank or waiting at an airport terminal for a flight to leave (Foxall & Soriano, 2005).

Accumulation in an open setting refers to saving and collecting (CC5), such as the

accumulation of coupons or other tokens before obtaining a product and instalments

payment for products or services that can only be taken when the full amount has been

paid (e.g. instalments for a holiday). Accumulation in a closed setting refers to token-

based consumption (CC6), such as air-mileages earned by airline frequent flyers, points


86

accumulated when staying at a hotel chain, and credit card points which consumers can

collect and redeem for a reward.

Hedonism in an open setting is described as popular entertainment (CC3), such as

watching a television show and reading novels, which provides hedonic rewards and

sensations. Hedonism in a closed setting refers to inescapable entertainment (CC4), such

as in-flight movies; although this behaviour is potentially pleasurable, it is still

unavoidable.

Finally, accomplishment in an open setting can be described as status consumption (CC1)

which consists of the purchase and consumption of status goods, such as luxuries and

radical innovations. Accomplishment in a closed setting refers to fulfilment (CC2). The

fulfilment category includes personal attainment (which has an element of recreation or

excitement) and personal achievements. Such behaviours are mainly maintained by social

rules because being a member of an exclusive social group means a necessity to conform

to its code of behaviour.

4.8.3. Level III: The Consumer Situation

To the question where is consumer behaviour? Foxall (1992b, p. 388) answers that

‘consumer behaviour is situated at the meeting place of the consumer and the setting’. It

is then the intersection of learning history (based on consumer experiences) with the

current setting scope that delineates a consumer situation (Foxall et al., 1998). It is

imperative to discriminate the consumer situation with the concept of behaviour setting

(Foxall, 1997a). A consumer behaviour setting acts as the discriminative stimulus which

signals a likely consequence when emitting a response. However, both the consumer and

the setting are essential for the definition of the consumer situation. It is the consumers’

learning history that determines what can act as discriminative stimuli and this history is


87

what accounts for the individuality of the consumer. It is the setting however that can then

activate the learning history and both can affect consumer behaviour (Foxall, 1992b). A

consumer situation should be seen as a particular (real world) setting along with a

consumer learning history (Foxall, 1997a).

4.9. Cognitive Psychology

Following the exploration of the Behavioural Perspective Model and moving to the next

of the two main approaches to study consumer behaviour, the study of cognitive processes

has resulted in the conception of consumer behaviour as a problem-solving and decision-

making activity the outcome of which depends on the way the buyer functions and on the

way the information provided to him/her is directed. The approaches proposed by

cognitive psychology have been well embraced by consumer behaviour research and have

added to knowledge creation in the field. Still, an extensive review of the cognitive

consumer research has showed that this approach to consumer behaviour is not without its

flaws and it has been increasingly criticised due to mixed findings (Foxall, 1990).

The initial use of extensive logical flow models of bounded rationality (e.g., Howard &

Sheth, 1969 as in Foxall, 1990, p. 10) has then deepened into what is often called the

‘information processing model of consumer behaviour’ (Bettman, 1979). The information

processing model regards the consumer as a logical thinker who solves problems to make

purchasing decisions. These problems are usually thought to be solved in a sequence of

steps involving such processes like stimuli, perception, motivation, memory, retrieval,

attitude change, intention and action. Although such models have been found to have an

application in cases of the involved and active consumer, they have been incapable to

accommodate and match all kinds of behaviours, as in the case of the uninvolved,

uninterested or sluggish consumer (Foxall, 1990). More simple, ‘peripheral routes’ (as in


88

the Elaboration Likelihood Model of persuasion- Petty & Cacioppo, 1986) have been

proposed for analogous cases. Such routes portray effortless and shorter processes that

consumers might follow when uninvolved or uninterested but the focus of such models is

on attitude change rather than actual behaviour. Others (see for example Kassarjian, 1978

as in Foxall, 1990) suggest that in these cases indeed simpler but preferably behaviouristic

models can capture actual consumer choice better.

The information processing perspective has further become widespread in consumer

research through the use of the notion of attitudes. Since its introduction into social

psychology (which both de Rosa, 1993 and Moliner & Tafani, 1997 accredit to Thomas &

Znaniecki, 1918), the concept of attitude has been defined in a variety of ways. In an

effort to sum up these definitions, attitudes can be considered as a three-dimensional

concept (having cognitive, emotional and behavioural dimension). The concept has been

used either as a mediating variable in many different theories of social psychology with

different meanings and applications or as a stand-alone examination of the attitudinal

process which can be faced as an evaluative activity (Eagly & Chaiken, 1993)– a concept

examining people’s opinions and beliefs.

Theories like the theory of reasoned action (TRA) (Fishbein & Ajzen, 1975), the theory of

planned behaviour (TPB) (Ajzen, 1991) or the technology acceptance model- TAM

(Davis, 1989) just to mention a few, are all ubiquitous examples of attitudinal theories.

The crux of such theories has been summarised by Howard (1983 as in Foxall, 1990) as

‘information-attitude-intention-purchase’ but the inconsistency between intentions to act

and actual behaviour has been established in past research (e.g. Davies et al., 2002). It is

also common among other trends in social psychology, like discourse and conversational

analysis (e.g. Potter & Wetherell, 1987), to critically debate the consistency of attitudes as

a stable organisation of dispositional traits, evaluative opinions, and especially plans of


89

action. This lack of empirical support for such a relationship between attitudes and

behaviour seems to be only overcome where measures of attitudes, cognitions or

responses reflected high levels of situational correspondence (Foxall et al., 2012, p. 462).

In any case attitudes possess both a heuristic role: they provide a simple strategy for

appraising an object and a schematic role: attitudes organise and guide complex behaviour

towards an object (de Rosa, 1993). Both strategies rely on a process that is impossible to

observe directly, because, as it is the norm in cognitive psychology, it is internal to the

subject. The observable part of the attitudinal process lies in the evaluative nature of the

responses a participant manifests about the object of the attitude (Moliner & Tafani,

1997).

Cognitive psychology is the standard form of understanding and doing consumer research.

Some of its basic theoretical and conceptual forms have however been found to be

inadequate to explain the nature of objects. The relevant concept of intentionality can be

proposed as a way to understand the nature of cognitive objects without resulting to strict

hypotheses like those accompanying the consumer as an information processing creature.

4.10. The Intentional Stance

The main characteristic of cognitive explanation is then the search for internal mental

states which can account for the observed responses of humans. This thesis is fundamental

to cognitive science with one specific philosophical notion, that of intentionality being

characterised according to Brentano as the ‘mark of the mental’ and the foundation to

cognitive psychology (Crane, 1998; Foxall, 2004; Crane, 2007). Still, in an attempt to de-

emphasise such sharp dichotomies, modern philosophy conceptualises intentionality as a

way to loosen the strict hypotheses on the nature of cognition that have been introduced


90

through cognitive psychology (Dennett, 2007) and Foxall (2007b) describes intentionality

as an alternative way to speak or explain the world.

Intentionality has been described as the power of the mind to be about, represent or stand

for things, properties or states of affairs. More importantly it is the characteristic of these

mental representations to have objects, to concern something, to be about something else.

However, although, many states like beliefs, desires, hopes, fears and anxiety are mental,

not all of these states can always be described as intentional (Searle, 2010). According to

Searle (1983) intentional states should be characterised by the following properties:

1. Intentionality should be perceived as directedness, thus only some and not all

mental states and events have intentionality. Namely, states like general anxiety,

stress, depression, confusion, when undirected and not towards specific

situations, objects or affairs cannot be perceived as intentional, however their

directed cases are intentional (Searle, 1983). On this same point Crane (2003)

adds that the objects of intentionality (intentional objects) can be of any kind-

ordinary objects, properties, events and states of affairs. One intentional state

might as well have many objects and some of them might not exist. He provides

the example of a belief that ‘my doctor smokes’. This mental state is directed

towards a person- the doctor, a state- smoking and a state of affairs- that the

doctor smokes; this belief can be true or not4. This property of intentional

language also known as intensional inexistence will be described in more detail

later, in the conceptual framework (chapter 6).

4 It is of interest to note that not all philosophers agree with this account of the nature of intentional objects.

Searle (1983, p. 17) in order to overcome the philosophical questions created when intentional objects do

not really exist proposes the distinction between the content of a belief (i.e. a proposition- that my doctor

smokes) with the objects of a belief (i.e. ordinary objects- in this case smoking and my doctor). Such

distinctions are mainly theoretical and are not perceived as having an influence to this analysis.


91

2. Secondly, intentionality is not the same as consciousness. There can be many

conscious states which are not intentional and many intentional states which are

not conscious. Thus according to Searle (1983), I might hold the intentional

belief for example that my ancestors spent their lives in a certain country but

until a point in time (for example when specifically asked about this) to have

never consciously formulated or considered this belief. Until this point in time,

my belief has been intentional but not conscious.

3. Intentions or intending to act, either in the form used by attitude theorists- as a

possible predictor of actual behaviour (see Ajzen, 1991), or in the form used by

moral theorists- that of acting intentionally or unintentionally (see Knobe, 2003)

should not be confused with intentionality. Although the obvious similarity

between the words would suggest a special role of acting intentions in the theory

of intentionality, according to Searle (1983, p. 3): ‘intending to do something is

just another form of intentionality along with beliefs, hopes, fears, desires’.

These three distinct uses of the word intention should not be misinterpreted.

4.11. Dennett’s Intentionality

A leading exponent of intentionality, Dennett, further describes it as a basic way to

analyse concepts like beliefs, desires, expectations, decisions and intentions. These terms

are used by folk psychology to predict behaviour of human beings or indeed computers

and animals (Dennett, 1983; Malle & Knobe; 1997; Dennett, 2007; Foxall, 2007a).

Namely, the intentional stance argues that any entity that its behaviour can be described

based on beliefs, desires or propositional attitudes forms an intentional system; this could

be a human being, a computer or a firm (Dennett, 1983; 1987).


92

Taking the intentional stance one step further Dennett’s (1987) intentional systems’ theory

claims that when explaining and predicting the behaviour of objects we can do that in

varying levels of abstraction. The more concrete the level the more accurate predictions

we can achieve. He thus describes the intentional stance, which assumes rational and

mentally capable agents as the most abstract strategy and he further introduces the

physical (the domain of physics and chemistry) and the design (the domain of mechanics)

stance (Dennett, 1987). Predicting the behaviour of a system based on the physical stance

assumes the knowledge of this system’s physical properties and the laws of physic that

define its operation. The knowledge of these parameters can predict the outcome of any

input with certainty. If the physical stance is difficult to be used, then the design stance is

an alternative way of prediction. The physical properties of an object need to be forgotten

here and one should concentrate on assumptions regarding its design. The design of every

object can give fairly accurate predictions that the system ‘will behave as it is designed to

behave under various circumstances’ (Dennett, 1987, p. 17). Dennett (1987, p. 17) sets

here the example of a computer. Most people are not aware of the physical properties of

the computer that contribute to it being a reliable and accurate machine. However, in case

they are aware of the functions that it is designed to do, they can easily predict the way the

machine will operate, leaving aside but at the same time being able to predict the

disconfirmation in cases of malfunctions. Although the design stance can extend to

biological objects like plants, animals and kidneys and hearts because these can be both

physical and designed systems, yet, only the designed behaviour of a system can be

expected from the design stance, abnormal behaviours are difficult to be predicted.

According to Dennett (1987) in cases when both these prediction strategies cannot work,

the intentional stance can then offer solutions. There are several steps which describe the

way the intentional stance works. First, the system of interest should be treated as a


93

rational agent; then the beliefs, desires, propositional attitudes of this agent should be

predicted based on its position in the world and finally assume that as a rational agent this

system will act in such a way that will comply to the furthering of its goals in light of

these beliefs.

In his most recent work, Dennett (2007) provides his explanation regarding the

relationship between cognitive psychology and intentionality by describing the intentional

stance as a ‘theory-neutral way’ of capturing the cognitive abilities of different organisms

without committing to exact hypotheses about the internal structures that underlie their

competences. This property of intentionality has found application to the development of

a more recent philosophical framework proposed by Foxall (2004; 2007a; 2007b; 2008)

called intentional behaviourism. As the name proposes, it draws and combines principles

of radical behaviourism and intentionality and suggests the use of intentional language in

order to overcome the limitations and extend the applications of behaviour analysis.

4.12. Intentional Behaviourism

To sum up, two main philosophical streams for explaining learning and behaviour have

been explored: radical behaviourism with its emphasis on operant conditioning and

cognitive psychology. Cognitive explanation seeks to understand the internal thought

processes and to explore their effect on behaviour. Intentionality has been described as the

power of the mind to be about, to represent (Crane, 2007), where systems are ascribed

thoughts directed at something other than them- the intentional stance. At the other end,

the main characteristic of radical behaviourism and operant conditioning is its avoidance

of intentional explanation and the use of a behavioural language which is based on

situational/ environmental influences- the contextual stance (Foxall, 2007a). These have

until recently been presented as incommensurable theories of behaviour.


94

Foxall, although an advocate of behaviour analysis and the proponent of radical

behaviourism in consumer behaviour (Foxall, 1986; 1987), argues that the explanations

provided by this stream of research is sufficient to predict behaviour in experimental

settings but when applied to real situations, it ultimately fails to give a complete

explanation of behaviour (Foxall, 2008a). Several aspects of human behaviour like the

personal level of explanation, the continuity/ discontinuity of behaviour and the

delimitation of human behaviour can better be explained by adopting intentional terms,

which can help to provide a more complete and accurate explanation of behaviour.

Consequently, the imperatives of intentionality (what the use of intentional terms adds to

behaviour analysis) according to Foxall (2007a; 2007b) are:

The personal level of explanation- the distinct way each person or group

experiences a situation based on their sensations and experiences.

The continuity of behaviour- an explanation on why a behaviour which is followed

by a particular reinforcing stimulus in a setting is re-enacted when encountering a

similar setting.

The delimitation of behavioural interpretation- the examination of open systems

rather than focusing on closed/ experimental settings only.

As a result, intentional behaviourism has been proposed (Foxall, 2004; 2007a and b) as a

way to accommodate both ways of thinking, a novel way to conduct research and

ultimately to facilitate the explanation of behaviour. Based on behaviour analysis and an

a-ontological conception of intentional states grounded on Dennett’s intentional stance,

new ways of conceiving and researching aspects of behaviour can be crafted which can

compensate for the shortcomings of the two aforementioned philosophical stances. It

compensates for the shortcomings of behaviour analysis, which due to the lack of


95

intentional explanation, results to the three aforementioned issues and those of cognitive

explanation which taken alone results to the de-contextualisation of human behaviour.

Intentional behaviourism thus draws attention to the necessity of employing intentional

mental language of beliefs and desires, intentions and propositional attitudes to account

for what is happening at the personal level of explanation, and hence invokes an

intentional explanation thereof. It can be used to explain operant behaviour in

experimental but more importantly in real settings where intentional explanation is

imperative.

In order to support and extend the interpretation of intentional behaviourism, two issues

related to the nature of intentionality and its role in the explanation of behaviour should be

clarified further. These two issues as deployed below are interrelated. The use of

intentional language as described by Foxall (2007a; 2007b) should be perceived as a

linguistic convention that carries with it no ontological implications regarding its nature

(Foxall, 2007a; 2008). Intentional objects hold then an a-ontological, linguistic nature.

Subjects are attributed the formation of verbal rules which are manifest when the expected

influence of contingencies is lost or altered (Foxall, 2008). However this relationship is

not enough to attribute causality to intentionality. It is merely to say that when an

individual’s actual rule-formulation coincides with the intentions we attribute to them,

their behaviour will be predictable in terms of behaviour analysis. The causes of the

behaviour are still to be found in the contingencies, though the questions (1) whether the

contingencies can consequently be modified by the person’s rule-making, (2) just how

initiating causes of overt and covert behaviour private stimuli are and 3) which are the

areas that the contextual and the intentional stance are both found to hold, remain to be

answered (Foxall, 2000; Foxall, 2008). The explanation of such behaviour and the answer

to these questions involve the ascription of intentionality and multiple theoretical and


96

empirical endeavours (Foxall, 2007b). According to most recent theoretical arguments

rules on the basis of rule-governed behaviour are constructs that can be used in

accordance with both models and areas where both explanations can find a simultaneous

application (Foxall, 2013; Oliveira-Castro, 2013). This faculty of rules to have a dual

nature will find application in the conceptual framework of this thesis.

4.13. Conclusion

This chapter has then provided the necessary knowledge foundations for the better

understanding of the propositions put forward by intentional behaviourism. It is this

knowledge base that can facilitate the understanding of the conceptual framework of this

thesis (chapter 6), which is based on the different approaches to exploring the Behavioural

Perspective Model. One of these approaches is based on an extensional understanding (the

contextual construal BPM-E) and the other on an intentional (BPM-I).

The subsequent and final chapter of literature review will focus on the description of the

emotional approach by Mehrabian and Russell (1974). The application of this model to

the examination of the principles of the BPM will be discussed. The variables of this

model will act as the main constructs for the present study but their explanatory power

will be based on their explanation and use as posited by the behavioural perspective

model.


97

5. MEHRABIAN AND RUSSELL’S ‘APPROACH TO

ENVIRONMENTAL PSYCHOLOGY’

5.1. Introduction

Through the exploration of the nature of confusion in chapter two, a brief but meaningful

examination of the main issues that the literature on emotions is negotiating along with

some theories of emotional ascription was achieved. This chapter will focus on the

emotional theory developed by Mehrabian and Russell (1974). The theory advocates the

effect of three main emotional responses to consumer situations, pleasure, arousal and

dominance which are said to mediate behavioural responses in the form of approach and

avoidance behaviour. The chapter will start by providing an overall evaluation of the use

of emotions in consumer behaviour and will then proceed to the examination of the

specific theoretical perspective. An account of the different approaches to its application

will be provided and the way this theory has been conceptualised in the boundaries of the

BPM will be discussed.

5.2. Categories of Emotional Theories and Consumer Behaviour

Although emotions have only recently become a subject of intense research (Barrett et. al.,

2007), due to the previous conception of individuals/ consumers as rational, information

processing creatures and the preoccupation with issues like memory, thinking and

personality (Holbrook & Hirschman, 1982; Foxall, 1992b), there is now a plethora of

theories and approaches that have been used in psychological inquiry, marketing and

consumer behaviour (O’Shaughnessy & O’Shaughnessy, 2003; Laros & Steenkamp,

2005). These approaches attempt to describe and measure emotional incidents.


98

A categorisation of theories relating to emotional ascription has been proposed in the past

(see Ellsworth & Scherer, 2003). This approach has grouped emotional theories widely

into three categories: the appraisal, the categorical and the dimensional theories of

emotions.

Appraisal theories describe emotions as adaptive responses to situations and very much

rely on individual interpretations of the world (Lazarus & Folkman, 1984). Theories

positioned within this category have been consistently describing consumer appraisals of

situations in cognitive terms (Lazarus et al., 1970). More recent developments in this

stream of theory have been promoting a less ‘causal’ and much more fluid relationship

between consumers’ cognitions (appraisals) and emotions (cognitive appraisals have been

described as possibly being all three- cause of emotions, part of the experience of

emotions and consequence of emotions- Roseman & Smith 2001), concepts like

‘emotionality’ have been introduced (Ellsworth & Scherer, 2003) and the idea that

appraisals are ‘hot’ rather than ‘cold’ cognitions have been accepted and supported by the

relevant literature (Lazarus, 1995).

Categorical theories, conceive the existence of some distinct, basic emotions (for this

reason these have been widely known as theories of ‘basic emotions’) or more recently

some families of basic emotions which are considered the result of evolution (Plutchik,

1980; Ekman, 1992; Izard, 2007). Such categorical theories posit that a limited number of

qualitative distinct primary/original emotions exist. According to that theoretical

argument, evolution has played an important role in shaping basic emotions’ unique

characteristics (Ekman, 1992). Plutchik's psycho-evolutionary theory is one of the most

influential classifications of general emotional responses. The theory proposes that the

following eight primary emotions -anger, fear, sadness, disgust, surprise, anticipation,

trust, and joy, should be considered as 'basic'. He further confers that these are the result

http://en.wikipedia.org/wiki/Anger

http://en.wikipedia.org/wiki/Fear

http://en.wikipedia.org/wiki/Sadness

http://en.wikipedia.org/wiki/Disgust

http://en.wikipedia.org/wiki/Surprise_(emotion)

http://en.wikipedia.org/wiki/Anticipation_(emotion)

http://en.wikipedia.org/wiki/Trust_(social_sciences)

http://en.wikipedia.org/wiki/Happiness


99

of evolution with an aim to increase the reproductive fitness of the animals, supporting in

this manner their biologically primitive nature (Plutchik, 1980).

Finally, dimensional theories posit that emotions, along with the affective qualities of

environments can be described along certain underlying dimensions such as pleasantness

and activation (Russell, 1980). Pleasure and activation are the most basic and widely used

dimensions, with the occasional addition of some other supplementary element like for

example dominance, potency or control (Osgood, 1966; Mehrabian & Russell, 1974;

Morgan & Heise, 1988). A final related approach to emotional categorisation has been

proposed to be the hierarchical approach. This approach specifies a hierarchical

structure in which specific emotions are particular instances of more general underlying

hierarchies, like for example the known hierarchy of positive and negative affect (Laros &

Steenkamp, 2005).

These theories differ mainly on the importance placed on the individual perception of

situations and the significance placed on certain emotions or affective qualities in relation

to others. For example, dimensional theories explain that all emotions can be characterised

by only some limited dimensions and in that sense can be summarised by them while

categorical place distinctive emphasis on the study of some distinct and specific basic

emotions (Mehrabian & Russell, 1974; Lazarus & Folkman, 1984; Ekman, 1992). All of

these approaches have found application in the study of consumer behaviour.

Laros and Steenkamp (2005) focused on the study of emotions in consumer behaviour and

compiled a table (see table 5.1) which indicates the diverse emotional approaches used by

consumer behaviour researchers.


100

Table 5.1 Overview of published consumer behaviour studies using emotions

as main variables

Reference Emotion measure used Resulting structure

Edell & Burke (1987) Edell & Burke (1987) Upbeat, negative and warm

Holbrook & Batra (1987) Holbrook & Batra (1987) Pleasure, arousal and

domination

Westbrook (1987) Izard (1977) Positive and negative affect

Olney et al., (1991) Mehrabian & Russell (1974) Pleasure and arousal

Holbrook & Gardner

(1993) Russell et al., (1989) Pleasure and arousal

Mano & Oliver, (1993) Watson et al., (1988)

Mano (1991)

Upbeat, negative, warm

Positive and negative affect

Oliver, (1993) Izard (1977) Positive and negative affect

Derbaix (1995) Derbaix (1995) Positive and negative affect

Steenkamp et al., (1996) Mehrabian & Russell (1974) Arousal

Nyer (1997) Shaver et al., 1987 Anger, joy, satisfaction and

sadness

Richins (1997) Richins (1997)

Anger, discontent, worry,

sadness, fear, shame, envy,

loneliness, romantic love,

love, peacefulness,

contentment, optimism, joy,

excitement and surprise

Dube & Morgan (1998) Watson et al., (1988) Positive and negative affect

Phillips & Baumgartner

(2002) Edell & Burke (1987) Positive and negative affect

Ruth et al., (2002) Shaver et al., (1987)

Love, happiness, pride,

gratitude, fear, anger,

sadness, guilt, uneasiness

and embarrassment.

Smith & Bolton (2002) Smith & Bolton (2002)

Anger, discontent,

disappointment, self-pity

and anxiety.

Verbeke & Bagozzi

(2003)

Frijda, Kuipers & ter Schure’s

(1989) Embarrassment

Yi & Baumgartner

(2004)

Appraisal theories (Folkman et.

al., 1986; Lazarus, 1991)

Anger, disappointment,

regret, worry. Source: Laros & Steenkamp, 2005 (with the addition of 2 recent studies)

The studies included in the table have had considerable influence in the consumer

behaviour realm. A closer examination of table 5.1 brings to the fore three key learnings

pertinent to the field:


101

1) All kinds of theories (dimensional, appraisal, hierarchical and categorical) have been

used in the exploration of consumer behaviour,

2) The use and measurement of hierarchical and dimensional categories of emotions (e.g.

positive and negative affect/ pleasure and arousal) (Westbrook, 1987; Olney et al., 1991)

are as common as the exploration of particular emotions like embarrassment, anger,

discontent and self-pity (Smith & Bolton, 2002; Verbeke & Bagozzi, 2003), and finally

3) The majority of studies successfully apply measurements of general psychology (Dude

& Morgan, 1998; Ruth et al., 2002). Emotional measurements specifically developed for

consumer situations or advertisement effects are not as widespread (for exceptions see

Edell & Burke, 1987; Richins, 1997).

This study will focus on the Mehrabian and Russell’s (1974) approach to environmental

psychology (and in that sense a dimensional approach to emotional environmental

ascription) in order to measure consumers’ reports towards descriptions of consumer

situations. This is the most concordant theory with the approach of this research. In

particular, this theory has found application in the study of the BPM, as it examines the

interaction of people with their environments and investigates the effect of environmental

learning and in this case, the levels of confusion, on consumers’ emotional reactions and

approach-avoidance behaviours.

5.3. Mehrabian and Russell’s Approach

Mehrabian and Russell’s approach is positioned within the broader field of environmental

psychology. Environmental psychology is an interdisciplinary field which is focused on

the interaction between human behaviour and its surroundings in many different settings

(Stokols, 1995). Early research in environmental psychology already widely

acknowledged the relationships between emotions–environments and behaviour–


102

environments (Berlyne, 1960). However, according to Mehrabian and Russell (1974)

initial research employed heterogeneous theoretical approaches and methods which could

not provide a sufficient structure to the study of human–environment relationships.

Mehrabian and Russell developed a more adequate framework of environmental

psychology and defined the field as one concerning two major topics: 1) the direct impact

of physical stimuli on human emotions and 2) the effect of the physical stimuli on a

variety of behaviours including social interactions, exploration and more importantly

approach-avoidance behaviours.

Mehrabian and Russell (1974) employed measures of three variables—pleasure, arousal

and dominance (PAD)—in order to describe and measure an individual’s affective

responses to an environment. Their theory argues that physical and social stimuli in any

environment directly influence the three dimensional view of emotions of individuals, and

consecutively affect a person’s approach and avoidance behaviour in it. Figure 5.1 depicts

Mehrabian and Russell’s approach to environmental psychology. This theory has become

widely known as a S-O-R (Stimulus-Organism-Response) approach (e.g. Mazaheri et al.,

2011). The original theory introduced the concept of information rate, the level of

information received by an environment or the ‘load’ of an environment, in order to

measure the environmental stimuli variable.


103

Figure 5.1 Mehrabian and Russell's approach to environmental psychology

Source: this study. Figure has been widely used to depict the theory by Mehrabian and Russell,

1974.

5.4. The Emotional Dimensions

Mehrabian and Russell introduced measures of three emotional dimensions, pleasure,

arousal and dominance. The measures of the three emotional dimensions are factorially

orthogonal (meaning independent from each other) so that any level of one of the three

dimensions can be accompanied by any level of the other two (Russell & Mehrabian,

1977; Foxall & Soriano, 2011). Within particular sets of stimuli these factors exhibit

however small, but significant, correlations and linear or curvilinear relations (Soriano et

al., 2013).

Pleasure is indicated by respondents’ verbal assessments of their responses to

environments as: happy as opposed to unhappy, pleased as opposed to annoyed, satisfied

as opposed to dissatisfied, contented as opposed to melancholic, hopeful as opposed to

despairing, relaxed as opposed to bored.

Arousal is verbally assessed at the extent to which respondents report feeling: stimulated

as opposed to relaxed, excited as opposed to calm, frenzied as opposed to sluggish, jittery

as opposed to dull, wide awake as opposed to sleepy, and aroused as opposed to

unaroused.


104

Finally, dominance is indicated by respondents’ reported feelings of being: controlling as

opposed to controlled, influential as opposed to influenced, in control as opposed to cared

for, important as opposed to awed, dominant as opposed to submissive and autonomous as

opposed to guided.

These PAD dimensions (along with the behavioural factors) are measured using

questionnaire based self-reported reactions to actual or descriptions of general or

consumer situations (Mehrabian & Russell, 1974; Russell & Mehrabian, 1978; Foxall,

2011). More specifically semantic differential scales are used, where all feeling states are

measured along single but bipolar dimensions. For example, contented-melancholic is a

feeling state ranging from extreme contentment to extreme melancholy, stimulation is

measured from feeling extremely stimulated to extremely relaxed and so on. The original

scale contains six adjectives to measure each emotional dimension. This classification is

respected and utilised in this research, as shown in table 5.2.


105

Table 5.2 The semantic differential measures of emotional states

Pleasure

happy–unhappy,

pleased–annoyed,

satisfied–unsatisfied,

contented–melancholic,

hopeful–despairing,

relaxed–bored.

Nine points scales

(e.g., extremely pleased to

extremely annoyed or extremely

happy to extremely unhappy).

Arousal

stimulated–relaxed,

excited–calm,

frenzied–sluggish,

jittery–dull,

wide awake–sleepy,

aroused–unaroused.

Dominance

controlling–controlled,

influential–influenced,

in control–cared for,

important–awed,

dominant–submissive,

autonomous–guided. Source: Mehrabian & Russell, 1974; Foxall, 1997b.

The advantage of the PAD scale is that it is simple and intuitive (Bagozzi et al., 1999).

However, when this approach is used to measure consumption-related emotions, several

limitations have been recognised. For example, Richins (1997) noted that because the

PAD scale was not designed to capture the entire domain of emotional experience,

researchers could not know the specific emotions, such as joy, guilt, or anger, being

experienced by customers merely from their PAD scores. Babin et al. (1998) also pointed

out that several researchers have uncovered distinct ‘positive’ and ‘negative’ emotion

factors while this approach which uses bipolar items might be inadequate for capturing

consumers’ positive and negative emotions simultaneously. This is problematic as these

same authors’ series of exploratory studies indicated that feeling a negative emotion does

not preclude the coexistence of a positive one. At the other end, Havlena and Holbrook

(1986) compared the Plutchik and the Mehrabian and Russell (M-R) schemes with respect


106

to their suitability for consumption activities. The results showed evidence in favour for

the latter, concluding that the three PAD dimensions captured more information about the

emotional character of the consumption experience than did Plutchik’s eight distinct

categories. The model has further been validated in many consumption settings and its

suitability to accurately describe consumers’ emotional responses, especially in context

specific settings, has been demonstrated (e.g. Foxall, 1997b).

5.5. Approach- Avoidance Behaviour

The three aforementioned emotional responses—pleasure, arousal, dominance—have

been proved to mediate overt consumer behaviours (approach–avoidance). The concept of

approach-avoidance has been the subject of research as one of the possible

conceptualisations of psychological coping. Coping mechanisms are adopted by lay

people in order to manage situations which can be classified as unpleasant, taxing or

exceeding the capacities of a person. The manifestations of coping have been studied

extensively within psychology (Lazarus & Folkman, 1984; Carver et al., 1989). Coping

has been considered an imperative and simultaneously complex psychological process

(Carver & Scheier, 1994). This complexity is indeed reflected in existing coping models

which illustrate that coping can be either problem or emotion focused (Lazarus &

Folkman, 1984), cognitive and behavioural (Carver et al., 1989) or as in the case of the

MR model an approach-avoidance motivational issue (Mehrabian & Russell, 1974).

The original behavioural measures (see table 5.3) of approach-avoidance included such

dimensions as relating to others, staying in or escaping from the setting, the desire to

explore and work in a situation and spending time there. More recently, and specifically in

consumer behaviour settings, more dimensions like the desire to consume or to spend


107

more money have been added (Mehrabian & Russell, 1974; Russell & Mehrabian, 1978;

Donovan & Rossiter, 1982).

Table 5.3 The four dimensions of Approach-Avoidance behaviour

Desire to stay

How much time would you like to

spend in this situation?

How much would you try to leave or

get out of this situation?

Desire to explore the situation

Once in this situation, how much

would you enjoy exploring around?

How much would you try to avoid

looking around or exploring this

situation?

Desire to work in the situation

To what extent is this situation a

good opportunity to think out some

difficult task you have been working

on?

How much would you dislike having

to work in this situation?

Desire to affiliate in the situation

To what extent in this situation

would you feel friendly and talkative

to a stranger who happens to be near

you?

Is this a situation in which you might

try to avoid other people, avoid

having to talk to them?

Source: Mehrabian & Russell, 1974 (original measures of Approach-Avoidance behaviour).

The theory was originally tested by Mehrabian and Russell (1974) with three

questionnaire experiments, where undergraduate students had to indicate their emotional

and behavioural responses to a varied set of described hypothetical situations, ranging

from being in a room listening to music to shopping in a grocery store. The results proved

that: 1) behaviours like exploration, affiliation and desire to stay in a situation are inter-

correlated and can be represented by the simple concepts of approach and avoidance

behaviour, 2) the emotional dimension of pleasure indicates the highest effect on

behavioural variables and 3) mixed results were identified as to the relationship between

arousal and approach which is said to be an inverted U-shaped relationship. In addition, an


108

interaction effect between pleasure and arousal in determining approach-avoidance

behaviour was identified (see also Soriano et al., 2013).

5.6. Applying the Mehrabian and Russell Approach

The Mehrabian and Russell (M-R) approach has been applied to consumer behaviour

research notably when situational effects of marketing are examined (Bitner, 1992). This

is very much in accordance with the theoretical basis of the PAD model which argues that

the impact of the situation on behaviour is mediated by emotional responses. The

application of the approach has however produced mixed and varied results. Lutz and

Kakkar (1975) did not find imposing results regarding the significance of situational

effects on consumer behaviour and contended for the inclusion of ‘other variables’—

possibly cognitive in nature— in the exploration of consumer behaviour so that the

situational approach does not remain isolated from other influences. It was Donovan and

Rossiter (1982) who re-introduced this environmental approach to consumer behaviour.

They measured the information rate (degree of complexity of an environment), the

emotional dimensions and approach-avoidance behaviour caused by several retail

environments.

Since then the approach has been applied in several contexts in order to measure all kinds

of environmental influences on consumers (atmospherics- Kotler, 1973); these include but

are not limited to external variables (buildings, architecture, surrounding area and

parking), interior environment (music, scent, temperature, lighting), store layout, interior

displays and human variables like crowding (as in Turley & Milliman, 2000).

More recently Walsh et al. (2011) have also attempted to portray a model with a

complexity of variables that could possibly influence consumers’ store satisfaction and

loyalty. They assumed that the relationship between store environmental cues (music and


109

aroma) - store choice criteria (merchandise quality, service quality and price) to

satisfaction and loyalty is mediated by the emotional variables of pleasure and arousal.

This study has then extended the exploration of the effect of pleasure and arousal to

additional variables than the typical approach-avoidance behaviour and has examined

more dimensions of consumer experience.

The results of the studies involving the MR approach are however varied and indicate that

at a basic level of explanation: 1) many variables can act as moderators when examining

such environmental influences and 2) results can be described as context and person

specific (Turley & Milliman, 2000). Regarding, for example, the effect of music it is

proved that the music played in a store has significant influence on emotional reactions

and several behavioural variables, like sales and time spent in the store, still this effect is

frequently moderated by variables like the age of the shopper (Yalch & Spangenberg,

1990 as in Turley & Milliman, 2000, p. 195), music tempo (Milliman, 1982 as in Turley

& Milliman, 2000, p. 195) and music preference (Herrington & Capella, 1996 as in Turley

& Milliman, 2000, p. 195). Moreover, varied results have been found as to the effects of

music on the emotional dimensions. Kellaris & Kent (1993) find a link between music

tempo and pleasure and arousal but Spangenberg et al., (2005) cannot find a direct link

between music and levels of arousal or pleasure. In addition, differences have been

reported in the effects of the emotional variables (especially dominance) on behavioural

measures with Donovan and Rossiter (1982) suggesting that the MR model is only a good

starting point for the exploration of consumer situations but more empirical work is

required for its reliable application.

Foxall (1997c) and Foxall and Greenley (1998; 1999) indicated that a possible reason for

these poor results has been that most studies use situations which are small in scope and

chosen arbitrarily. Studies based on the Behavioural Perspective Model have indicated


110

more consistent results regarding the applicability of the PAD model to varied

descriptions of consumer situations. These situations are formed based on the levels of

utilitarian and informational reinforcement and the closeness or openness of the setting as

described by the BPM and have been consistently been found to be explained on the basis

of the variables of the MR model (Foxall, 1997b; 1997c; Foxall & Greenley; 1998; 1999;

2000; Foxall & Soriano, 2005; Abu Hasan, 2011).

Specifically, pleasure has been described as an index of the utilitarian reinforcement

signalled by the situations or by the usage of products and services implicated. This is so

because utilitarian reinforcement consists of the benefits and satisfaction contingent in a

situation. Arousal is a measure of the informational reinforcement which indicates the

feedback on consumer performance and finally, dominance is predicted to increase with

the degree of openness of the behavioural setting. Thus consumers are expected to feel

more controlling influential and important in an open rather than a closed setting (Foxall

&Soriano, 2005). These theoretical explanations have resulted to the following complete

BPM contingency matrix (figure 5.2) where levels of the PAD are indicated.


111

Figure 5.2 The BPM Contingency Matrix

BEHAVIOUR SETTING SCOPE

Closed Open

ACCOMPLISHMENT

(high utilitarian, high

informational)

Contingency

Category 2

Fulfilment

+P +A –D

Contingency

Category 1

Status Consumption

+P +A +D

HEDONISM

(high utilitarian, low

informational)

Contingency

Category 4

Inescapable

Entertainment/Pleasure

+P -A –D

Contingency

Category 3

Popular Entertainment

+P -A +D

ACCUMULATION

(low utilitarian, high

informational)

Contingency

Category 6

Token-Based Consumption

-P +A –D

Contingency

Category 5

Saving and Collecting

-P +A +D

MAINTENANCE

(low utilitarian, low

informational)

Contingency

Category 8

Mandatory Consumption

-P -A –D

Contingency

Category 7

Routine Purchasing

-P -A +D Source: Foxall, 1992a and b

Regarding the behavioural measures, approach behaviour is expected to increase with the

total quantity and quality of reinforcement (utilitarian or informational) while avoidance is

expected for lower levels of reinforcement. It is accepted that people will have the

tendency to approach situations that offer more and better reinforcements and will avoid

those that lack such reinforcing qualities. In addition, approach minus avoidance

behaviour (a variable named aminusa and which will be described in more detail in the

methodology chapter) is also expected to increase with the openness of the setting. In

simple terms, consumers are predicted to have a tendency to approach open situations and

avoid closed ones.


112

This examination of the BPM until now uses ‘within-the-skin events’ (Skinner, 1974) like

emotions and approach- avoidance behaviour and argues for the empirical interest of such

terms. In this case, emotional and behavioural variables represent the emphasis of radical

behaviourism on the use of language as an indication of verbal behaviour. This language

no matter if it is overt and public (as in conversations) or covert and private (as in

thinking), it can be representative of behaviour (Foxall, 1990; Foxall; 1998; Foxall et al.,

1998). Thus verbal behaviour in the way that has been used to explore the BPM until

recently should be understood as plain statement of the facts and a description of its

functional relationships with environmental events, and particularly the relationship with

its contingencies and consequences.

5.7. A Role for Dominance

Extending on the research implicating the emotional dimensions of Mehrabian and

Russell (1974), one final clarification is essential regarding the emotional dimension of

dominance. Dominance is the least researched of all 3 PAD components in retail

environments (Soriano & Foxall, 2006).

Russell (1978; 1979) and Russell & Pratt (1980) first claimed that, without the dominance

dimension, only pleasure and arousal can adequately capture the emotional reactions to

stimuli. They based their argument on empirical and theoretical grounds. They claimed

that pleasure and arousal account for large proportions of variance of the affective quality

of environments while dominance has a minor role and can be perceived as a secondary

and possibly cognitive dimension. Namely, dominance might denote beliefs about

consequences and antecedents of the emotional states and in that sense requires cognitive

intervention. This conception of dominance is more in accordance with appraisal theories

of emotions (e.g., Roseman, 1984; 1991; Smith & Ellsworth, 1985; Lazarus, 1991) which


113

argue for the cognitive antecedents of emotions and less in accordance with a theory

searching for the affective qualities of environments (Russell & Pratt, 1980).

This claim has however gained partial support by empirical studies in consumer behaviour

(Donovan & Rossiter, 1982). Donovan and Rossiter (1982) were the first to indicate that

dominance might not have an effect on behaviour in diverse shopping environments. They

gave justice to their results based on Russell and Pratt’s (1980) work and acted as the

milestone for future retail research, which based on Donovan and Rossiter (1982)

removed dominance from their theoretical argument and measuring instruments (e.g.

Baker et al., 1992; Donovan et al., 1994; Mattila & Wirtz, 2001; Lee et al., 2011; Walsh et

al., 2011).

Although the use of only two factors (pleasure and arousal) is enticing due to greater

simplicity, at the other end, important distinctions between emotions can only be

accomplished by using all three emotional dimensions (Mehrabian, 1996 cited in Soriano

& Foxall, 2006). Soriano & Foxall (2006) support this argument and meticulously indicate

the importance of dominance in consumer behaviour. Utilising valid arguments it is

indicated that dominance has been mistreated in consumer behaviour research, since

Donovan and Rossiter (1982) replaced some of the items of dominance with random

terms and did not use the original scale, and has been removed on arbitrary reasons and

without enough and clear evidence. Dominance can consistently discriminate between

closed and open environments according to the classification of the Behavioural

Perspective Model (Foxall & Greenley, 2000; Soriano et al., 2002) and Ward and Barnes

(2001) also proved that at least in a particular environment, that of fast food restaurants,

dominance has a direct effect on affect, store involvement, attitude and behaviour. Recent

articles that utilise the bi-dimensional emotional approach by using only pleasure and

arousal also point to the importance of further tests on the effect of dominance in


114

retail/choice environments in order to further test/validate its applicability to these

contexts (Walsh et al., 2011).

Finally, consumers have been found to connect confusing situations with feelings of

‘helplessness’ and ‘being overpowered’ (Plutchik, 1994; Schweizer, 2004) and perceived

lack of control has been identified as a central appraisal leading to confusion. On these

grounds, dominance will be included in this study.

5.8. Conclusion

Although the theory of Mehrabian and Russell (1974) is not designed to capture specific

and distinct emotions and it was initially intended to describe affective reactions to

diverse environments (including but not limited to consumer situations), the approach has

been widely applied to consumer environments. The results of these studies are not

definite and leave much space for further exploration for example in terms of the

situations used, the interactions among the emotional terms and especially, the role of

dominance in consumer settings.

The main strength of the approach, as noted by Lutz and Kakkar (1975) and successfully

utilised by the BPM research programme, is that its implementation allows not only the

description of distinct conditions but also facilitates the comparison of different situations.

Since all situations can be described based on the level of pleasure, arousal, dominance

and approach-avoidance behaviour that these produce to individuals then a comparison of

diverse situations is feasible based on the analysis of the relevant scores. Then it is easy to

examine situational differences based on either the personal (groups of consumers with

similar characteristics) or the aggregate level (the summative scores produced for the

situation) and reach interesting conclusions about the qualities of such settings.


115

Specifically Mehrabian and Russell‘s affective and behavioural measures as verbal

responses have been used in previous research as psychometric measures of the

consumers‘ verbal behaviours in specific situations (Foxall, 1997b; Foxall & Greenley,

1999; 2000; Foxall & Soriano, 2005). The test of the BPM (Behavioural Perspective

Model) has been done through the prediction of consumers’ verbal responses to

descriptions of specific consumer situations. This research programme that linked a

consumer‘s verbal behaviour of his or her emotional reaction to consumer situations has

clearly supported the link between verbal reports and BPM’s prediction of consumer

behaviour (Foxall, 1997b; Foxall & Soriano, 2005).

In the case of this research, verbal behaviours are expected to arise in response to the

interaction between verbal descriptions of the situated consumer behaviour and the

respondents’ history of reinforcement or punishment in similar situations.

CHAPTER 6- CONCEPTUAL FRAMEWORK

116

6. CONCEPTUAL FRAMEWORK

6.1. Introduction

The preceding four chapters have been an overall evaluation and presentation of the extant

literature. These have been an attempt to present and justify the necessity for this novel

proposition on the nature of confusion and simultaneously lay the foundations for the

theoretical underpinnings of this study. Chapter six on the conceptual framework connects

relevant arguments in order to provide the overarching theoretical framework that will

guide this knowledge inquiry. It starts by providing a briefing on the state of confusion

and proceeds by re-examining the current state of the Behavioural Perspective Model

(BPM) as a device using an extensional language (BPM-E). It will continue by placing the

model within the framework and study of an intentional explanation (BPM-I). Confusion

can be defined in terms of a self-based rule (or better a rule for the lack of rules). There is

however a differing language that can be used to describe confusion– the first,

extensional, deals with confusion as an overall response to physical and social stimuli and

the other, intentional, deals with it in terms of human intentionality (as in Foxall, 2013, p.

118). The explanatory or interpretative role that the construct can play in these models

will be described. Specific research hypotheses that correspond to these explanations are

developed. Overall, this conceptual framework supports the supposition put forward by

Foxall (2004; 2007a; 2007b; 2013) that when the extensional understanding becomes

exhausted we need to turn to intentional language in order to expand our understanding;

both of these approaches are offered to researchers to facilitate their endeavours.


117

6.2. The State of Confusion

In an attempt to summarise previous chapters and although a consensus on the state of

confusion has not been reached, most psychology and consumer behaviour researchers

agree that confusion as a state can be meaningfully characterised by the following

qualities (Ellsworth, 2003; Hess, 2003; Keltner & Shiota, 2003; Rozin & Cohen, 2003a

and 2003b; Schweizer, 2004; Walsh & Mitchell, 2007; 2010):

A state of not knowing/ understanding.

A sense of goal obstruction which in consumer behaviour might equal to either an

inability to choose the preferred/ best product or to the impediment of an enjoyable

shopping.

Perceived higher levels of effort, higher attention needed and possibly a sense of

lack of control.

Intense uncertainty and/or impressions of overload, similarity, novelty etc.,

especially operationalised as such in consumer research.

These characteristics are depicted5 in table 6.1.

Table 6.1 A depiction of the main characteristics (‘qualia’) of the state of

confusion

A state of not knowing/ lacking understanding

Sense of goal

obstruction

Inability to choose/

enjoy the shopping

experience

Ambiguity/ Similarity/ Overload.

Variety/ Novelty/ Complexity/ Conflict/

Comfort/ Reliability

Higher levels of perceived effort and attention necessary

Source: this study (based on the characteristics attributed to confusion in previous theoretical and

empirical papers).

5 In light of the fact that diagrams are usually perceived by readers as representing a ‘flow’ and possibly

cause and effect relationships, a diagrammatic presentation of the qualia of confusion as described in

previours research has been avoided in this instance. Rather ‘qualia’ is the subjective experience, the

‘overall feeling’ of a situation and this idea should not be reduced to impressions of causal relationships. A

table which mearly signifies the characteristics has been used instead.


118

Based on the literature as reviewed in chapters 2 and 3, the theoretical opportunities

offered through the study of states like confusion have been established (Rozin & Cohen,

2003).

Confusion has been further described as a pure emotion, a pure cognition, a combination

of the two (cognitive and emotional situation) or a cognitive feeling and also as either a

context and time specific response to an environment or as a personality proneness (like

the general psychological tendencies established in psychology). There is also extended

debate on the differences between the conscious and subconscious parts of confusion,

with conscious parts being described as the easier to capture and measure.

This study will extend the intellectual interest for the study of confusion based on a

concept deriving from behavioural psychology; the idea of rule-governed behaviour and

specifically self-based rules.

6.3. Confusion as a Self-Based Rule

The concept and especially the importance of rule-governed behaviour (other or self

instructed) for the study of behaviour have been described before (refer to chapter 4).

Rules are usually defined by social norms however self-rules are dictated when the

speaker and the listener are the same, thus are dictated by the self. Self-rules act as

instructions and are effective as long as they adhere to norms to which an organism has

followed throughout its history (Foxall, 1997a). Based on the categories of rules

developed by Zettle & Hayes, (1982), especially the case of tracking is concerned with

corresponding to a description of the state of affairs (Zettle & Hayes, 1982, p. 79–92) or

according to Foxall, 2013 (p. 118) it is a case of ‘responding to brute facts’ like the

arrangement of the physical environment. The arrangement of the physical environment

indicates the state of affairs as the consumer is usually powerless to change it and needs to


119

adhere to that. Tracking can be viewed as predominantly a contingency-shaped behaviour

and although most theorists (Zettle & Hayes, 1982; Foxall, 1997b; Törneke et al., 2008;

Foxall, 2013) argue that it is a challenging task to clearly define and understand the

different cases of rule-based behaviour and discern among plys, tracks and augments,

confusion can be understood as a case of a self-based track due to its special relationship

with environmental conditions.

6.4. Confusion as ‘Anomy’

In an attempt to further the above understanding, the concept of anomy (or ‘anomie’) will

be brought to the fore. In its true meaning anomy comes from the Greek language and it

means the absence of law. The concept of anomy was initially introduced by the French

sociologist Émile Durkheim and subsequently deeply analysed by Merton (as in

McClosky & Schaar, 1965). In sociology Durkheim used the term to describe a state of

normlessness, deregulation and loss of social control usually produced by too sudden

social change. Merton extended the concept to indicate (Merton, 1938; also Merton, 1957

as analysed in Lowe & Damankos, 1968) that this deregulation is the result of the Western

(USA) society’s increasing emphasis on accumulation of wealth which is not

accompanied by the relevant emphasis on the means to obtain these monetary goals. This

is causing strain to the relevant social groups that do not have the means to attain the

goals, leading to their isolation. In sociology, anomy is then a characteristic of social

groups whose access to goals is blocked by social-structural barriers. Merton’s approach

to anomy is reputed as the pre-eminent sociological theory of deviant behaviour.

In psychological research anomy has been portrayed as a state of mind rather than a state

of the society or social groups (McClosky & Schaar, 1965). It has mainly been described

in terms of the alienation and dis-institutionalisation of the individual from others, the


120

society or the goals that the social system approves, and it is usually measured by a

relevant scale and conceptualisation developed by J. L. Srole (Merton, 1957; Taylor,

1968). According to a more general definition and approach which is focused on an even

less sociological and increasingly psychological perspective (as in McClosky & Schaar,

1965, p. 19):

‘anomy is a state of mind...it is the feeling that the world and oneself are adrift,

wandering, lacking in clear rules and stable moorings...for him (for the individual) the

norms governing behaviour are weak, ambiguous and remote.’

Anomy, simply defined, is a rule for the lacking of rules; it is a state where norms or rules

are confused, unclear (complexity/ ambiguity confusion) or absent (similarity confusion)

and learning of the norms is severely impeded due to all of these reasons (McClosky &

Schaar, 1965). The case of confusion seems to correspond to this kind of reasoning.

Different kinds of confusion can be characterised by the lack of market rules and norms

which interfere with learning and impede behaviour. Confusion can be characterised

intensely by the sense of market anomy, this sense of disorientation, which can be defined

as a rule characterising the lack of other relevant rules.

6.5. Confusion as a Self-Based Rule (or ‘A Rule for the Lack of Rules’)

Confusion can then be perceived as a case of self-tracking (self-based rule) and more

specifically a ‘rule’ suitable to describe the lack of other relevant rules. The role of self-

rules as summarised by Zettle & Hayes (1982) has been to ‘being personal’ in the sense

that a person can react more effectively now or in a future occasion than when based on

the contingencies alone.

Extending further on this theoretical reasoning, a fundamental faculty of rule-governed

behaviour according to Foxall (2013) is the capacity of being treated and expressed in


121

both extensional and intentional terms. This logic follows Searle (as in Foxall, 2013) who

concludes that items can be perceived in both an extensional account of ‘brute facts’ and

an intentional based on ‘human intentionality’. The meaning and application of this

principle in the case of confusion will be described in the following sections. This study

will then extend the understanding of confusion by placing it within the framework of

extensional and intentional BPM and thus a novel understanding of the construct and the

application of the BPM will be offered.

6.6. The Languages of Explanation

Having described confusion as a self-based rule (or better the lack of rules that can guide

behaviour) and examining rules’ unique characteristic to be described in terms of either an

extensional or an intentional language it is essential to re-examine what these languages

represent.

Lay people very often use language which attributes actions and intentions to other

individuals’ desires and beliefs, and both people (and researchers as already explained)

widely use this approach in order to understand and often predict human behaviour.

Behavioural science on the other end deals with such an approach with circumspect due to

the ease with which explanations of any behaviour can be adduced by assuming that goals

and dispositions from the behaviour they are said to explain are used to explain that same

behaviour (Foxall, 2013). This has resulted in reaching a state resembling the chicken and

the egg situation. On these grounds, the preferred approach for the investigation of the

BPM until recently has been the use of an extensional language (in terms of simple verbal

behaviour), the avoidance of intentional or cognitive terms and at the same time the

determination of the explanations that the extensional language can provide by observing

the inadequacies of the intentional stance.


122

It must be clear by now that these ways of expressing reality correspond to the contextual

and the intentional stance previously discussed. This conceptual framework chapter

concerns and will expand on the ways that the principles of intentional behaviourism can

be applied to the exploration of consumer choice when the main device of exploration is

the BPM and part of this endeavour is the understanding of the different languages that

can be used to explore this model.

More specifically, the most important characteristic of extensional language is simply that

it avoids intentional terms. In this kind of reasoning ‘a stimulus is a part of the

environment which is consistently followed by a response’ (Foxall, 2013, p. 108) and the

idea that an organism expects, believes or desires something do not have a role in this

explanation. At the other end, the intentional explanation exists exactly at the level of

personal beliefs and desires and embodies terms that refer or represent something other

than themselves.

In order to properly mark the difference between the two languages, the defining

characteristics that distinguish between the two can be exemplified as following:

1. The extensional language is characterised by referential transparency while

intentional idioms are referentially opaque. Referential transparency means that in

any extensional sentence synonymous terms can be used to substitute one another

without changing the value or meaning of the sentence. This is not valid in

intentional idioms. One example used to indicate this property (Foxall, 2007a and

b; Foxall, 2013) is the sentence ‘That planet is Mars’. In this extensional use of the

language ‘Mars’ can be easily substituted by the ‘forth planet from the sun’.

However when saying, ‘John believes that this planet is Mars’ Mars cannot be

substituted by ‘the forth planet from the sun’ simply because John might not know


123

or believe that mars is the forth planet from the sun and thus by substitution the

meaning of the sentence might completely change and lose its original meaning.

2. Intentional language is characterised by intensional inexistence (while extensional

language by physical existence). This characteristic of intentional language has

been described before (see also chapter 4). It means that an intentional sentence

does not imply its true existence or non-existence. When an extensional sentence

states that ‘John bought a BMW’ this implies that John and a car brand named

BMW exist. However, an intentional explanation which argues that: ‘George

thinks that John bought a BMW’ does not imply the existence of either the action

or the brand itself. This belief is inside the individual and it is not necessarily

positioned in the actual word.

3. Finally, according to Brentano (1874/ 1973 as in Foxall, 2013) and based on both

the above characteristics it is difficult to translate intentional into extensional

sentences. However according to Searle items and constructs can be identified and

described in the extensional -the physical level- but also in accordance to human

intentionality. According to Foxall, 2013 (p. 118) rule-governed behaviour (in the

form of tracks, plys and augments- Zettle & Hayes, 1982) are behaviours that carry

this property. Rule-governed behaviour (directed by self or other rules) can be

explained both as responses to social and physical stimuli and as ideas expressed

in accordance with human intentionality.

On these grounds and the overall understanding of the two accounts it is safe to argue that

intentionality does not provide the same kind of description as an extensional explanation,

however it can be used to explain behaviour whenever an extensional language no longer

suffices (Foxall, 2007a; 2007b; 2013). This is usually when the continuity or discontinuity

of behaviour, the personal level of explanation and the delimitation of behavioural


124

interpretation is sought (Foxall, 2004). In this manner two distinct models are proposed,

the extensional (BPM-E) and the intentional model (BPM-I).

6.7. The Behavioural Perspective Model Extensional– (BPM-E)

The benefits of using the behavioural perspective model have been described in previous

research (Foxall, 1992a and all subsequent research). This model which has been until

recently depicted in the extensional language of stimuli, behaviour and behavioural

consequences (lacking reference to beliefs, desires or other intentional attitudes) has been

introduced as an attempt to overcome the limitations of the cognitive portrayal of choice,

especially the de-contextualisation of theoretical models. To start with, the model offers a

relevant framework which accommodates the two main reinforcers innate in consumer

situations, utilitarian and informational. It further describes decision-making with respect

to settings that range from the routine, habitual and everyday to the extreme. When using

the model several diverse situations have been described and categorised based on the

reinforcers and behavioural setting. It places distinctive emphasis on the idea of consumer

situation, which is the way that behaviour is located in space and time by the extensional

model. Figure 6.1 depicts the model and explains its main constructs (see also chapter 4).


125

Figure 6.1 BPM-E

The Extensional Behavioural Perspective Model. The variables are extensionally defined

as responses to physical and social stimuli embedded in the consumer situation. The

consumer situation (coterminous with the consumer behaviour setting scope) consists of

the consumer behaviour setting (discriminative stimulus, motivating operations and verbal

rules) and the learning history. Reinforcement is comprised of Utilitarian (UtilR) and

Informational (InfR). Punishing or aversive consequences are also part of the possible

consequences, conceptualised and examined in previous research mainly in terms of the

cost of buying.

Source: Foxall, 2013, p. 110.

Following the extensive previous literature these constructs have been conceptualised as

the Pleasure (UtilR), Arousal (InfR), Dominance (consumer situation/ behaviour setting

scope) and the approach/avoidance behaviour as the consumer behaviour element. The

way these have been used in the extensional construct has been in terms of overall

responses to stimuli. These represent then verbal behaviour (in accordance with Skinner,

1974) expressed with the help of an extensional language and should not be perceived as

representing consumer’s beliefs of attitudes. The language of radical behaviourism is very

specific in this occasion and it is very strict on the role of a discriminative stimulus. Thus

a discriminative stimulus does not represent or signal utilitarian and informational

reinforcers or punishers it simply ‘sets the occasion’ for them (Foxall, 2013, p. 111). It

Consumer situation

Consumer behaviour

Consequences

Reinforcement

Consumer Behaviour Setting*Learning History

UtilRInfR


126

neither allows for personal or group differences, disregarding in this manner the personal

level of explanation.

In the boundaries of the BPM-E confusion can then be defined as a rule-governed

behaviour (tracking) which is a ‘response to the physical and social environment’. It is an

aversive, extensionally defined- objective consequence of environmental exposure to

specific discriminative stimuli/ markets. The role of the UtilR, InfR and confusion (which

can have both UtilP and InfP implications) can facilitate the examination and

establishment of overall differences in stimuli means. Figure 6.2 below adds confusion to

the BPM-E.

Figure 6.2 BPM-E

The addition of confusion is depicted in this revised model.

Source: this study (Foxall, 2013, p. 110 the model has been revised to depict the extensional

understanding of confusion as a response to a stimulus).

It is also relevant to argue that for the purposes of this research (which deals with the

nature, effect and addition of confusion in the BPM) the whole situational complexity of

the BPM cannot find application. Consumers do not hold confusion for the range of

operant classes (accomplishment, hedonism, accumulation and maintenance) and

situations previously described by the model. Situations like being in a job-related

Consumer situation

Consumer behaviour

Consequences

ReinforcementPunishment


UtilRInfRUtilP (unishment)InfP (unishment)

Confusion


127

seminar, driving an expensive car, being on a cruise or collecting loyalty card points

(these are all situations used to describe the contingency categories of the BPM in

previous research) are inappropriate for this research. Such situations have been

specifically chosen and manipulated in previous research mainly in order to establish the

measurement of the Mehrabian and Russell variables as good indicators for the aspects of

the behavioural perspective model.

In order to achieve the explanation of the contextual and intentional stance in the case of

confusion other specific choice/shopping related situations had to be used. Thus this study

maintains the basic premises of the model on the importance of the reinforcers, the

behaviour setting scope (proved to be conceptualised and measured in past research as the

pleasure-arousal-dominance variables of the Mehrabian and Russell model) and

approach/avoidance behaviour and will extend the basic principles of the BPM beyond the

original model in different shopping situations where confusion is expected to pose an

effect. In that sense this study will allow for a free exploration of differing consumer

situations and will not predetermine/manipulate but only hypothesise the expected levels

of the extensional value of the different variables.

A main limitation of the BPM as examined until now is that the model has only been

tested in terms of reinforcements- utilitarian (pleasure) and informational (arousal).

However, the effect of aversive consequences although depicted in the original BPM (as a

line connecting utilitarian and informational reinforcement with aversive consequences)

has been examined mainly as the effect of monetary cost, which is indeed one of the main

aversive consequences of consumer choices (Sigurdsson et al., 2010). Confusion can be

described as self-based rule which in an extensional language is translated into ‘an

aversive consequence/punishment of shopping’ and the extent of its effect needs to be

examined.


128

An additional main limitation of this ‘extensional’ approach and conception of the BPM

and specifically confusion is that it removes the personal level of explanation (the level of

personal rules in the form of beliefs/ propositional attitudes) from the understanding

developed. In order to examine this personal level, behaviour should be reconstructed and

discussed in terms of an intentional account. This account takes into consideration not

only the environmental effect but also the consumer’s perception of shopping, what the

consumer has been led to believe in terms of their own experiences of other similar or not

situations and actually what he/she desires. It is then possible that a consumer might find a

complex environment as more acceptable than another consumer who based on a previous

experience was not able to buy the desired product based on unavailability. This consumer

will act differently to the variety of products on offer than another consumer with different

perception and experience. In this case of the personal level of explanation, we have no

other resort than to turn to the language of intentionality, the language of beliefs and

desires. By adopting the intentional language or stance we adopt a ‘less scientific’

approach to the study of phenomena; but since the social world lacks the comfort of

constant experimental conditions, where the complexity of the learning history of objects

can be known, social scientists need to resort to such language in order to better explore

phenomena (Foxall, 2013).

It is on these grounds that the inclusion of intentional terms and finally an intentional

conception of the BPM (BPM-I) is proposed (Foxall, 2004; 2007a; 2007b; 2013) so that

these shortcomings of the extensional BPM can be faced. The flexibility that Searle offers

on the multiple uses of terms allow us to employ the same constructs in both the

extensional and intentional way and this rule will be followed here in order to indicate the

ways that an intentional BPM can add to the understanding provided by the extensional

BPM.


129

6.8. The Behavioural Perspective Model Intentional– (BPM-I)

In order to prove and describe the value of the Intentional Behavioural Perspective Model

(BPM-I), Foxall (2013) describes the addition of the concept of collective intentionality

and explains the implications of this collective understanding for the model. This study

will extend on this idea but contrary, the individual intentionality in the form of a tracking

self-based rule (confusion) will be examined in order to incorporate the personal level of

explanation to the model. Figure 6.3 depicts the intentional BPM.

Figure 6.3 BPM-I

The Intentional Behavioural Perspective Model. The central explanatory component of

the BPM, the consumer situation is redefined in this new understanding. Consumer

Situation in this intentional model can be found ‘in the complex of the representation

and meaning which intentional construals’ supply (Foxall, 2013, p. 107). Behaviour is

then transformed from reactions to presented stimuli into intentionality-directed

behaviour.

Source: Foxall, 2013, p. 116.

Rules and rule-governed behaviour are capable of contributing to the two kinds of

explanations as described above. These can be described in both the extensional sense, as

stimuli that come to have the same effect as non-verbal contingencies that can act to

predict behaviour. In addition, the alternative intentional explanation treats rules as

Consumer

situation

Consumer

behaviourConsequences

Reinforcement

Intentionality

Beliefs

Propositional Attitudes

Desires

UtilR

InfR


130

representations of the three term contingencies that act at the personal level and can be

described using the language of beliefs, the intentional language.

On these grounds and as one of the least proposed ideas in the psychology and consumer

behaviour realm confusion can also have the characteristics of an intentional state

meaning that it can act at the belief/ propositional attitude level. At this level confusion

can be described as:

‘An individual belief about something else —in the case of consumer research— a

personal belief that a specific market is confusing’.

Although it is critical to acknowledge a) that some instances of confusion can be

objectless (as described by Searle, 1983), meaning that there can be cases that one is

difficult to describe what one is confused about and b) that some personality types have

the tendency to get (or describe oneself as getting) more easily confused than others; this

study advocates that confusion can have the properties of an intentional state meaning it is

directed towards specific objects which in that instance are particular markets/shopping

situations.

The adoption of the intentional stance in this case can theoretically provide a solution to

the two important debates on the construct which derive from previous psychological and

consumer behaviour theory and research, namely the everlasting cognition-affect debate

and the one related to the conscious- subconscious nature of confusion.


131

6.8.1. The Debate on the Cognitive-Affective (or Binary) Nature of

Confusion

Based on the previous discussion on intentionality6, this philosophical concept has been

described as proposing the loosening of the ties of the strict hypotheses guiding cognitive

psychology. It does not need to resort to the language of internal mental or information

processes. Intentionality expresses the ability of the mind to stand about something else,

which usually takes the form of beliefs/ attitudes/ desires and which are part of an

individual experience and learning history in connection with the environments it deals

with. On these grounds, distinctions and debates on the cognitive and affective nature of

constructs are not an integral part of intentional explanation. Following these arguments

the exact nature of constructs, principally in terms of cognition/ affect is not

epistemologically necessary in this kind of explanation.

6.8.2. The Conscious-Subconscious Debate

Confusion can be described as an intentional belief/ propositional attitude held by

consumers. In this proposed intentional conception of confusion the distinction between

conscious and subconscious confusion is again an argument without real value. As

described by Searle (1983) an intentional construct might be subconscious (consumers

might perceive a market as confusing without ever realising it) but this can come to the

fore only when consumers are actually asked specifically about it.

Intentional behaviourism maintains that ‘intentional ascription in terms of beliefs,

attitudes, desires is rational for an individual to have in view of a specific situation

defined by the intersection of his/her learning history and the setting he/she faces’ (Foxall,

6 Refer to chapter 4.


132

2007, p. 43). It is then reasonable to suggest that intentional confusion can be a kind of

propositional attitude which takes the form of a developed self-based rule about specific

shopping environments (thus it takes the form of track) that in accordance with intentional

behaviourism is the result of previous consumers’ experiences with the setting (the result

of a learning history) along with individual/personal differences, which in this explanation

can signal the consumer situation. This learning history can result to consequences in

cases consumers are to behave within the specific setting (which acts as the discriminative

stimulus) in the future. It is imperative to clarify that based on the above definition and the

nature of intentionality, intentionality is ascribed to the individual7 (Foxall et al., 2012, p.

477) and is developed through consumers’ idiosyncratic experiences and characteristics.

In this manner not all consumers face environments in the same way and nor do all

consumers have the same learning histories in a setting. This is exactly what has been

described as ‘the personal level of explanation’ by intentional behaviourism. In this

manner not all consumers are holding the intentional confusion in specific settings. The

level of reported confusion varies based on the specific characteristics of the setting and

especially individual understanding. Consumers can then be arrayed on a mental

continuum which extends between confused and not confused consumers in each setting.

In this study the distinction between confused and non-confused consumers will be treated

as dichotomous (due to the nature of statistical tests, like ANOVA, which requires

splitting variables in order to implement hypothesis tesing); however the continuous

nature of confusion should be kept in mind when interpreting the results. Figure 6.4

depicts the inclusion of confusion in the intentional BPM.

7 For an exception of this idea see the concept of collective intentionality (e.g. Searle, 1990), which is

attributed to collective entities.


133

Figure 6.4 BPM-I

The Intentional Behavioural Perspective Model based on the understanding provided by

rule-governed behaviour. Behaviour is transformed from reactions to presented stimuli

into intentionality-directed behaviour.

Source: this study (as in Foxall, 2013, p. 116 figure has been revised for the requirements of this

study).

Foxall (2000, p. 777) describes that the contextual stance can and should be used

whenever the intentional stance is used because the juxtaposition of opposing arguments

can help knowledge to grow.

The reason for using both the intentional and the contextual stance is that the observed

behaviour can be fully explained only by ascribing it to both stances (Foxall, 2009, p.

219). It is then evident that by ascribing beliefs and attitudes, the personal level of

explanation and the way that works in specific markets can be explained. However it is

also relevant to argue that the intentional and contextual stance can interact to produce an

effect on behaviour and this aspect will also be examined in this study, mainly drawing on

the concept of experience.

6.9. Main Proposition of this Thesis

Based on the underpinnings of the above theoretical understanding, the main proposition

suggested by this study is:

Consumer

situation

Consumer


Reinforcement

Punishment

Intentionality

Belief/ Propositional Attitude = Individual Confusion

UtilR

InfRBehavioural Setting Scope


134

Confusion is a self-based rule (based on the propositions of rule-governed behaviour). It is

more specifically, a rule for the lack of other rules (a case of market anomy). Due to its

relationship with the state of affairs (environmental situations), it can be characterised as a

self-based track and as such it can be treated at two levels.

At the extensional level, it can be treated as a response to specific (discriminative) stimuli

and can act along with verbal contingencies to predict behaviour. In this case it represents

verbal behaviour; it is a plain statement of the facts.

At the intentional level, it is the result of the interplay between an individual and specific

situations and in this case it can take the role of the consumer situation that signals

consumer responses. By adopting this ‘less scientific’ route, it can be assumed that

confusion can have an impact on actual situational contingencies. Such an approach based

on intentionality allows for the personal level of explanation to be examined.

Table 6.2 The nature of confusion at the two levels (extensional and

intentional) proposed by this study

BPM-E

Lack of (complex, weak or similar) market rules or norms that

impede behaviour; measured as plain facts and overall responses

to consumer situations.

BPM-I Individual perception of the lack of (complex, weak or similar)

market rules or norms that impede personal behaviour. Source: this study

When levels of intentional confusion are accompanied by a further lack of other

self/personal rules that can guide decisions and choices then confusion is expected to have

a stronger effect on emotional and behavioural aspects in this situation. This thesis will

then attempt to combine and investigate a possible combined effect of the contextual

(situational effects) with the intentional stance (see group of hypothesis 2).


135

6.10. Development of Research Hypotheses

The following examination of the hypotheses will commence by building on the

intentional understanding of the differences between confused and non-confused

consumers. Based on this level of analysis the choice of the specific retail situations

examined will be justified and the chapter will proceed to the extensional description of

the BPM and the hypotheses guided by this approach.

6.10.1. Group of Hypotheses 1

Foxall & Soriano, 2011 acted as the first attempt to establish Roll’s proposition (Roll,

2000; 2005) that ‘individual differences influence both conditionality and emotionality’

via hypothesis based on the theory of adaptive-innovative cognitive style (Kirton, 1976 as

in Foxall & Soriano, 2011). Their study used the eight situations proposed by the BPM

and was based on the fact that adoption-innovation as a cognitive style correlates

positively with such characteristics as extraversion-introversion, flexibility, tolerance of

ambiguity, self-esteem and sensation-seeking (as in Foxall & Soriano, 2011). They

propose then that consumers characterised as innovators should prefer relatively open

settings and in these settings they should report higher PAD and approach scores than

adaptors. At the other end, adaptors should prefer relatively closed settings. The BPM

indicated once again its robustness as a theoretical model, however no relationships with

the cognitive variables (adoption-innovation) as described were identified (Foxall &

Soriano, 2011). This is one indication of the way previous research has integrated the use

of behavioural and cognitive approaches.

In the context of this research and considering the character of confusion as a self-based

rule, markets act as the external agents that provide the environment and are supposed to

provide the norms of shopping. Success or failure depends upon the progress of the


136

individual. Reinforcement is provided by understanding the rules and performing the act

while punishment by failing. Confusion is then understood as an aversive consequence/

punishment of consumption and shopping and for those consumers who hold the

intentional confusion, there will be personal changes inside the contingencies of the

market (expressing in that sense the personal level of explanation), measured as a

difference in the behavioural variables and consequently to emotional (reinforcement).

Specifically, it has been indicated (Foxall, 2010b, p. 328) that the physical contingencies

can be influenced by verbally based contingencies that can have the power to augment,

decrease or in cases even replace the naturally-occurring reward system. Following the

flow of figure 6.4, confusion is expected to influence consumer behaviour in specific

settings and this should then have an effect on the utilitarian and informational

reinforcement that is experienced in such settings. It should be kept in mind that the two

situations in this case are perceived as either highly confusing or non-confusing.

a. Behavioural Variables

Due to the nature of confusion as an aversive consequence/punishment, approach is

expected to be lower for confused consumers while higher levels of avoidance are

expected when compared to non-confused consumers:

Hypothesis 1: Overall, the range of confused consumers will indicate lower levels of

approach behaviour than the range of non-confused consumers.

Hypothesis 2: Overall, the range of confused consumers will indicate higher levels of

avoidance behaviour than the range of non-confused consumers.


aminusa (approach-avoidance) behaviour than the range of non-confused consumers.


137

b. Pleasure/ Utilitarian Consequences

Confusion has been connected in past research with affective consequences and coping

mechanisms. Utilitarian consequences have been described as the functional results which

are usually mediated by both the situation but also the buying and usage of products

themselves. One of the reasons confusion is related to situational affect is because it is

connected to the consumers’ judgement of whether the environment will facilitate or

frustrate goal achievement (Ellsworth, 2003). Pleasantness is defined as the hedonic

valence (pleasant or unpleasant) of the affective response to a stimulus; it is based on the

extent to which the stimulus (the object of the affective response) enables people to

achieve their salient goals. Namely in the case of situations that are perceived as

facilitating goal achievement (perceived as non confusing) these engender positive affect

and are experienced as having higher utilitarian reinforcement. When environments are

perceived as impeding goal achievement (by confused consumers) these evoke the

opposite effects and are experienced as offering lower utilitarian reinforcement (Kaltcheva

& Weitz, 2006; Ward & Barnes 2001; for a full review Clore et al., 1994). Regarding the

case of product buying and usage, it is also expected that confused consumers will feel

less pleased overall with the products they buy due to the feelings of uncertainty that the

buying process has developed. Confusion has been proven to have a significant negative

impact on consumers’ overall macro satisfaction with situations (Walsh & Mitchell,

2010).


pleasure (utilitarian reinforcement) than the range of non-confused consumers.

c. Arousal/ Informational Consequences

Moving to the topic of informational reinforcement, the discussion becomes more

complicated. Drawing upon information theory, Mehrabian and Russell (1974) have


138

demonstrated that the arousing quality of an environment correlates highly with its

information rate, which increases with the novelty complexity, intensity, unfamiliarity,

improbability, change, mobility or uncertainty of the setting- characteristics that have

been conceptualised and measured as confusion. However, Russell and Mehrabian’s

(1997) second study of the article ‘Evidence for a three factor theory of emotions’ found

significant evidence that confusion can be described as an unpleasant (mean= -0.53),

unaroused (mean= 0.27) and submissive state (mean= -0.32), means ranging from -1 to 1

(see table 4 of the aforementioned article—case 121— confusion). It is then a

controversial matter whether the aversive consequence of confusion influence consumers’

arousal levels or not and one that has scarcely been examined by previous research.

In light of the proposal of the BPM that arousal levels indicate and measure the

informational value of an environment (described as feedback on personal performance

usually mediated by social rules and described as indicating social status) an alternative

theoretical explanation for the relationship between confusion and arousal will be

proposed. The theoretical propositions of the BPM conceptualise arousal mainly as a

measure of the ‘symbolic power’ such as social status and self esteem of situations and

this symbolic power is what provides the feedback on performance. It will then be of great

value to test whether a different kind of feedback on performance, the one provided by

confusion, which can be perceived as a feedback on the level of understanding/

‘cleverness’ of the consumer, which might result to a decrease in self-esteem will actually

have an effect on arousal.

In case consumers perceive confusion as their personal incompetence this will have a

negative effect on the perceived levels of personal performance, thus:


139

Hypothesis 5a: Overall, the range of confused consumers will indicate lower levels of

arousal (informational reinforcement) than the level of non-confused consumers.

This study will also keep a more open approach on the relationship between arousal and

confusion, because according to the principles of the BPM there are chances that arousal

measures status/ symbolic feedback on performance only and not the kind proposed by

confusion. Thus, the possibility that the null hypothesis of 5a (that there is no relationship

between the two variables) has a theoretical meaning will be considered:

Hypothesis 5b: Overall, the range of confused consumers will indicate the same levels of

arousal (informational reinforcement) with the range of non-confused consumers.

It is interesting to note that based on the theoretical grounds of this study (mainly the

proposition of the BPM that arousal is a measure of the feedback on performance) the

relationship between confusion and arousal can be either negative or lacking (in

agreement with Russell and Mehrabian, 1997 who identified that confusion is a relative

un-aroused situation) but not positive in the way that the relationship between information

rate and arousal has been described in previous research. It should also be kept in mind

that Donovan and Rossiter’s study (1982, p. 54) reported that three information rate

measures (novelty, density and size) had a positive relationship with arousal, variety a

negative one, while irregularity had no significant relationship with it. All of these

findings are then an indication of a relationship that requires further clarification.

d. Dominance/Behaviour Setting Scope

In terms of the third and least researched emotional element of the PAD one should draw

upon the special meaning of dominance (as a measurement of the scope of the behaviour

setting- openness or closeness of situations) in the BPM. It can be said that confused

consumers will feel more that their choices are dictated by agents outside themselves (as


140

they themselves are incapable of reaching decisions or consider markets to be outside

their understanding) and on these grounds will perceive environments as more closed.

Confusion has anyway been connected to feelings of helplessness and being overpowered

as described previously (Schweizer, 2004). Following this logic:


dominance than the range of non-confused consumers.

As described by the previous handling of the BPM, behaviour would also be expected to

increase with the total quantity and quality of reinforcement available to reinforce it

(Foxall & Soriano, 2005). It has also been expected that Approach–Avoidance scores for

open consumer behaviour settings will significantly exceed those for closed settings.

Aminusa (the overall behavioural evaluation) is then expected to be lower for confused

consumers for those two additional reasons as described above. In confusing ‘situations’

the overall levels of reinforcement (utilitarian/informational) are expected to be lower and

these ‘situations’ are also expected to be perceived as more closed (refer above).

This first group of hypothesis describes the personal level of explanation as put forward

by intentional behaviourism. It is here expected and hypothesised without giving a true

ontological nature and using it only as a way to understand reality that personal rules

developed by individuals in the form of intentionality will have an effect on the

contingencies of the markets, because confusion represents an environmental punishment

in any environment.


According to Foxall (1997a, p. 105-106), the levels of experience characterising a

situation regulate decision-making in this context. When a market is characterised by high

levels of experience, it is perceived as more habitual and consumers eventually develop


141

more tracks, which are akin to the rules identified by Zettle & Hayes (1982). They will

then use these tracks in order to reach easier decisions and get easier to a particular goal

point. In this manner, it is expected that the effect of confusion, as a self-based rule, will

be less in the market with the higher overall levels of experience.

Table 6.3 Behavioural and cognitive approaches to decision-making.

Low experience High experience

BPM

Other rules. Consumers

lack a relevant learning

history-prompt search for

other rules- external or

internal to the individual.

Self rules. Acquisition of a

learning history, from which self

rules can be extracted.

Elaboration

Likelihood Model Central route Peripheral route

Mode Deliberation Spontaneity

Heuristic-systematic

processing Systematic processing Heuristic processing

Source: Foxall, 1997a; Foxall, 2000.

More specifically, when the learning history in a market is such that known consequences

have followed regularly and unimpeded from specific acts, the discriminative stimuli in

the setting will provide signals that more quickly result in the performance of the requisite

behaviour. In accordance with this logic, when the market has become a routine the actual

contingencies in this habitual market are stronger and consumers on average are expected

to have already developed a repertoire of rules and norms to act upon (Foxall, 1997, p.

105-106; also Foxall, 2000). Experience within a market represents the result of a stronger

previous learning history in this situation. Thus, an environmentally determined context

(we are turning here to an extensional language again), is expected to act as a kind of

moderator in the relationships among confusion- PAD and A_A. This relationship seems

to be especially relevant to consumer confusion. Confusion is a construct strongly related

to understanding and anomy, thus it seems very concordant to argue that the decreased

levels of experience in a market can increase its expected effect in that market.


142

Thus the following hypotheses have been developed to depict the relationships expected:

Hypothesis 7: The effect of confusion on pleasure will be stronger for the market


Hypothesis 8: The effect of confusion on arousal will be stronger for the market


Hypothesis 9: The effect of confusion on dominance will be stronger for the market


Hypothesis 10: The effect of confusion on approach behaviour will be stronger for the

market characterised by overall lower levels of experience.

Hypothesis 11: The effect of confusion on avoidance will be stronger for the market


Hypothesis 12: The effect of confusion on aminusa (approach-avoidance) will be stronger

for the market characterised by overall lower levels of experience.

This group of hypotheses examines the way that the contextual and the intentional stance

interact and indicate whether the effect of confusion on the emotional and behavioural

variables depends on the situational influence of the markets examined.

6.10.3. Choice of Situations Examined

Based on the above two groups of hypothesis the relevant situations used in this study

need to adhere to the following criteria:

Situations that confusion is expected to be an issue and have an effect; in that

sense mainly choice/shopping situations should be used.


143

Situations that could be agreed (based on theoretical and practical arguments) to

differ in terms of expected levels of experience.

Several situations and markets have been examined when it comes to confusion as

explained in chapter 3. In addition, several situations were considered by this study when

examining the situations that could fit the above criteria. The final choice was to perform

a comparison between two markets that could fit the above criteria, namely: the grocery

market and the buying of high technology products. Confusion has been described as been

a problem in both situations (Friedman, 1966; Khermouch, 1994; Cahill, 1995; Leek &

Chansawatkit, 2006; Schweizer et al., 2006). In addition, the level of familiarity and

experience is expected to differ in the two markets for multiple reasons.

Grocery shopping has been described and used in previous research in cases that ‘a

routine buying situation’ (Baharrel & Denison, 1995) is sought for, while buying of high

technology products has been used when ‘exciting- novel’ and innovative situations

(Parasuraman & Colby, 2001; Lee et al., 2011) are examined as part of a research. In

terms of shopping frequency it is also expected that consumers are on average much more

familiar with grocery shopping compared to high technology products’ shopping, as

grocery shopping is a much more frequent and every day activity.

To finish with, even the stresses and hassles examined in the literature are associated with

not knowing how to make decisions or deal with product characteristics when examining

high-technology products (Mick & Fournier, 1998). Anxieties and stresses described for

grocery shopping (Aylott & Mitchell, 1999) are of an everyday and procedural nature, like

parking availability and trolley shortage.

Based on these arguments grocery shopping and high-technology shopping have been the

preferred situations for this project. In order to facilitate understanding and achieve better


144

focus the shopping of high technology products has been represented by the description of

PC/Laptop shopping. The levels of consumer experience with each market have been

measured based on an overall-macro evaluation of experience and secondarily based on

the reported frequency of shopping; the results of both of these measurements are reported

later in the analysis.

The following hypotheses will now turn to the extensional understanding and the

developed hypotheses will correspond to that kind of language:


The approach taken to examine the BPM has established the use of the MR model as a

useful and adequate tool in order to compare situations based on the levels of utilitarian

and informational reinforcement, levels of closeness or openness (e.g. Foxall & Soriano,

2005). In the case of the BPM-E the discriminative stimuli are the descriptions of the

situations and the contingencies acting in the situations are what cause the differences in

the levels of the variables. In the case of this study the discriminative stimuli (SD) are the

two chosen markets/choice situations. Grocery shopping has been used in past studies of

the BPM as a description of the situation used to depict the Contingency Category 7

described as ‘Routine Purchasing’ (relatively open setting scope- compared to the rest of

the situations in the model). It is further placed within the wider operant class of

‘Maintenance’, based on low utilitarian and low informational reinforcement compared to

the other categories of the model. It is then worthwhile to examine the relevant position of

a High Technology purchasing situation into the categories and classes of the BPM. This

answer seems difficult because conceptually (and following close examination of the

other categories of the BPM model- Fulfilment/ Status Consumption/ Inescapable

Entertainment/ Popular Entertainment/ Token-based Consumption/ Saving and Collecting/

Mandatory Consumption) it only seems to belong to the same category and operant class


145

as grocery shopping. It is however evident that the two situations are not exactly similar,

several characteristics of the two markets differ. In addition, although high technology

products are more and more everyday devices, PC/Laptop in itself might equally have

some characteristics of hedonism (pleasure seeking), considering the value of high

technology products as passing-time/ entertainment appliances/ gadgets. It is then

worthwhile to wonder whether situations that conceptually seem to belong in the same

contingency category and operant class differ significantly in their levels of

reinforcement. Based on theoretical, practical and empirical arguments hypotheses on the

comparison of the levels of utilitarian, informational, closeness-openness, confusion and

behavioural responses of the two markets will be developed.

However in accordance with this study’s exploratory nature, data will finally reveal

whether differences can be expected in the case of such situations. However based on the

nature of the products, purchases and situations involved the expected relationships will

be described.

a. Pleasure/ Utilitarian Reinforcement

As utilitarian consequences indicate the value-in-use, the economic, pragmatic or material

consequences of purchases (Foxall et al., 2006, p. 103)

Hypothesis 13: The two markets are expected to differ in terms of utilitarian

reinforcement with the high technology market expected to have higher pleasure than the

grocery market.

b. Arousal/Informational Reinforcement

Hypothesis 14: The two markets are expected to differ in terms of informational

reinforcement with the high technology market expected to have higher arousal than the

grocery market.


146

c. Dominance/ Behaviour Setting Scope

The two markets chosen for this endeavour are both expected to be open situations and

consequently to indicate higher levels of dominance since the modern retail competition

allows for what has been previously described as ‘consumer democracy’ (Lane, 2000;

Schweizer et al., 2006). Consumer democracy is expressing the attempt of all markets to

place consumer in the centre of all choices and offer freedom of choice by providing all

the necessary products. However it can be reasoned that due to the fact that grocery

market is more of an everyday situation consumers will be freer to act (external agents

will have a lesser influence upon their decisions) within this market. Thus:

Hypothesis 15: The two markets are expected to differ in terms of dominance with the

high technology market expected to have lower dominance than the grocery market.

d. Confusion

Being in the choice situations like the ones proposed in this study entails some punishing

utilitarian and informational consequences. Any situation entails both punishing and

reinforcing consequences as described by the BPM. For instance the utilitarian

reinforcement of being in an exotic holiday has to do with the rest and relaxation it can

offer. This type of holiday might equally produce punishing utilitarian consequences like

the cost of buying the trip and a long and tiring flight. Equally informational

reinforcement involves the prestige and sense of accomplishment, however punishment

might also derive from negative social comments that this kind of trip is snobbish and

culturally meaningless (example as in Foxall et al., 2006, 103-104). In the same manner

one of the aversive consequences of choice situations are levels of confusion

(environmental anomy) that consumers are exposed to and have to deal with. As described

before confusion can possibly have both utilitarian and informational consequences and

these consequences can determine behaviour.


147

At the levels of the comparison regarding levels of confusion in the two markets, there is

not enough evidence to argue for or against any possible comparison. The following

hypothesis will be left open ended (post-hoc comparison) and allow for the data to

indicate the final relationship.

Hypothesis 16: Possible differences are expected in the levels of confusion in the two

markets that act as discriminate stimuli in this study.

e. Behavioural Variables

This study will further test the assumption laid forward by the BPM that behaviour will be

expected to increase with the total quantity and quality of reinforcement available to

reinforce it. Thus:

Hypothesis 17: Approach will be higher in the market characterised by higher levels of

utilitarian and informational reinforcement (thus the high technology purchasing situation

is expected to have higher approach).

Hypothesis 18: Avoidance will be higher in the market characterised by lower levels of

utilitarian and information reinforcement (thus the grocery market is expected to have

higher avoidance).

Hypothesis 19: Aminusa, the net difference between approach and avoidance will be

higher in the market characterised by higher levels of utilitarian and informational

reinforcement (thus the high technology purchasing situation is expected to have higher

levels of aminusa).

Thus the levels of pleasure/arousal/ dominance/ A_A and the comparison between the

markets will depend on the discriminative stimuli which in this case are the descriptions

of the situations.


148


This study will finally test its main theoretical argument that pleasure, arousal, dominance

and confusion will determine approach-avoidance consumer behaviour. The two situations

have been chosen in order to represent pure shopping/choice situations and in this manner

a re-examination of the robustness of the MR model in consumer environments will be

achieved. More importantly, the capacity of confusion as a valid construct that can work

along with the PAD variables to add value to the BPM is to be examined. Based on the

literature on the PAD, the BPM and the propositions of this study the following

hypotheses are proposed:

Hypothesis 20: Affective variables of pleasure, arousal and dominance will each have a

positive relationship with approach. Confusion will have a negative relationship with

approach.


negative relationship with avoidance. Confusion will have a positive relationship with

avoidance.


positive relationship with aminusa, the net difference between approach and avoidance.

Confusion will have a negative relationship with aminusa.

Hypothesis 23: Aminusa (the net difference between approach-avoidance) will be

determined by the variables pleasure, arousal, dominance and confusion.

Hypothesis 24: Two-way interactions may be identified between the affective variables

pleasure and arousal (possibly dominance) when considering their effect on aminusa.


149

6.11. Conclusion

This chapter has laid the foundations and achieved the description of the conceptual

framework that this research is based on. The exploration of the theoretical underpinnings

and the findings of previous research regarding confusion, rule-governed behaviour, the

BPM, and the applications of intentional behaviourism in the extension of the BPM to an

intentional understanding (BPM-E/ BPM-I) resulted in the extraction of important

information, based on which the proposed framework of this research was structured and

the hypotheses were developed.

An understanding of confusion based on the principles of rule-governed behaviour has

been offered and the implications for the conception and study of confusion have been

discussed. Essentially confusion has been described as contributing to an extensional and

an intentional exploration of the BPM and the way these ideas can drive theoretical

development and research practice into new directions are presented.

The Mehrabian and Russell (1974) model has been described as possessing the faculty to

measure the main variables of the BPM model, utilitarian and informational

reinforcement, behaviour setting scope and approach-avoidance behaviour. The capacity

of these measurements to represent such ideas has been proven in previous research

(Foxall, 1997b).

Leaving aside the main propositions of the proposed frameworks (the BPM-E and the

BPM-I), the issue still remains that very limited research has examined in a systematic

manner the level and effect of confusion on differing consumer situations and the

relationship with emotional and behavioural responses, notably concerning the variables

proposed by the MR model; this will act as a supplementary objective of this study.


150

On these grounds, one bi-directional main proposition (using an extensional and

intentional language) and several research hypotheses based on theoretical grounds and

concepts have been described. For the purposes of facilitating this knowledge inquiry, an

appropriate research methodology has been utilised and will be deployed in the

subsequent chapter.

CHAPTER 7- RESEARCH METHODOLOGY

151

7. RESEARCH METHODOLOGY

7.1. Introduction

Many researchers still remain very much preoccupied with the choice between

quantitative and qualitative methods. However, issues related to research conduct go far

beyond research methods because the choice of research methods is just one of the four

elements of research methodology. The others are: the scientific research paradigm,

identification of the research design and also the data analysis techniques (Saunders et

al., 2009, p. 108). Chapter seven will be organised around those four major topics of

methodology. It explains the position of this study in relation to the scientific research

paradigms. It further describes the research design and methods used to collect and

analyse the data allowing for the exploration of the research questions and hypotheses that

were previously proposed. The criteria to establish and judge research integrity and

quality will also be discussed.

The issues around research methodology are of decisive importance for every research

because these have a threefold goal: to determine what can be accepted as knowledge in a

research project and to legitimise that knowledge within an acceptable framework of

processes (Benton & Craig, 2001). To finish with, it dictates the way the methods and the

analysis of findings will be conducted.

7.2. Scientific Research Paradigm

The examination of the scientific research paradigm of this thesis will commence with an

exploration of established philosophical frameworks for conducting research. Based on a

critical examination of these philosophical frameworks, two basic beliefs that guide the

overall conduct of this research will be established. The section will then extend to the


152

explicit philosophical beliefs of this study in terms of ontology, epistemology and

methodology.

7.2.1. Overview of Research Paradigms and Guiding Beliefs of this Inquiry

Scientific research paradigms are overarching conceptual frameworks within which some

researchers work, that is, a paradigm is a world-view or a set of linked assumptions about

research conduct which is shared by a community of scientists investigating the world

(Kuhn, 1962 as in Deshpande, 1983, p. 101). Philosophical paradigms act as an attempt to

organise the confused and contradictory world of common sense and provide guidance to

research (Foxall, 1995). It has been conventional since Kuhn (1962) to call these

particular combinations of assumptions paradigms. Kuhn’s work in the natural sciences

presupposed that paradigms generally succeeded one another. Contrary, in the social

sciences, it has been admitted that a set of antithetical paradigms can exist simultaneously

(Mingers, 2001).

Specifically for the broad area of marketing and consumer behaviour research, Pachauri

(2002) argues (see also Shankar & Patterson, 2001 on the same topic) that these have been

generally characterised by two broader paradigms, positivist (emphasising the economic,

behavioural, cognitive, motivational, trait, attitudinal perspectives)—which is dominant

since the late 50’s, giving a culturally neutral ‘etic’ account (see also Foxall, 1995;

Goulding, 1999) and non-positivist (interpretive and post-modern perspectives—

emphasising the symbolic, subjective experience), which emerged during 1980’s

exploring a culture-specific ‘emic’ account. Peter and Olson (1983, p. 120) have

summarised the defining characteristics of these broader approaches where empiricist

science is determined by the discovery of the true nature of reality, and relativism by the

more ‘loose’ approach, where science is allowed to create many versions of reality.


153

One of the most cited tables (and book chapters) illuminating different paradigms in social

sciences is the one by Guba & Lincoln (1994, p. 109) — table adapted here based on

Perry et al., 1997 as in Healy & Perry, 2000 and Guba & Lincoln, 2005). Table 7.1

distinguishes then four major research paradigms: positivism, post-positivism, critical

theory and interpretivism/constructivism, and explains them in terms of their defining

elements: ontology, epistemology and methodology.

Table 7.1 Four social research paradigms explained

Positivism

Post positivism

(realism) Critical Theory

Interpretivism/

Constructivism

Ontology

Naive realism- reality

is real and

apprehensible. Reality

is driven by

immutable natural

laws and its true

nature can only be

obtained by testing

theories about objects,

processes or structures

in the real world.

Reality is real

but only

imperfectly and

probabilistically

apprehensible.

Virtual reality

shaped by social,

economic, ethnic,

cultural and

gender values,

crystallized over

time. It has also

been named

historical realism

as social reality is

perceived as

historically

constituted.

Relativism:

Multiple, local and

specific constructed

realities. The social

world is produced

and reinforced by

human through their

action and

interaction.

Epistemology

Objectivist: findings

are true. Verification

of hypotheses through

rigorous empirical

testing, search for

universal laws. Tight

coupling among

explanations,

predictions and

control.

Modified

objectivist:

findings

probably true.

Transactional/

subjectivist: value

mediated

findings.

Knowledge is

grounded in

social and

historical

practices;

knowledge is

generated/

justified by a

critical evaluation

of social systems.

Transactional/

subjectivist;

understanding of the

social world through

the participants

perspective; through

interpretation of

their meanings and

actions; researchers

come into research

with previous

assumptions.

Methods

Hypothetical-

deductive

experiments/

manipulative;

hypotheses testing;

chiefly quantitative

methods.

Case studies/

convergent

interviewing:

triangulation,

interpretation of

research issues

by qualitative

and by

Dialogic/

dialectical;

researcher as a

‘transformative

intellectual,

changes the social

world within

which participants

Hermeneutical/

dialectical:

researcher as a

passionate

participant within

the world.

Interpretive case

study, action


154

quantitative

methods.

leave. Critical

ethnography;

interpretive case

study; action

research.

research, holistic

ethnography.

Source: Based on Guba & Lincoln (1994, p. 109); Perry et al., 1997 as in Healy and Perry, 2000;

Guba & Lincoln, 2005.

A paradigm is thus a construct that specifies a general set of philosophical assumptions

covering ontology–the form and nature of reality. This is a term used to describe the

answer to the question: ‘What kind of things are there in the world?’ (Benton & Craib,

2001), epistemology–the relationship between the researcher and what can be known, the

nature of valid knowledge and the methods–the techniques and practices used by the

researcher to investigate that reality (Krauss, 2005) .

The less influential but very relevant for the scopes of every research project is

praxiology. Praxiology defines the way that researchers act in an informed and reflective

manner. According to Habermas (1993) (as in Mingers & Brocklesby, 1997) praxiology is

subdivided into research effectiveness (a. the relevance of research questions to the

paradigm employed and b. practical issues such as how trained a researcher is on specific

methods and way of thinking) and ethics (also called axiology).

It has then been described that the established paradigm in consumer behaviour has been

the ‘traditional’ positivist philosophies of science (Shankar & Patterson, 2001; Pachauri,

2002). These philosophies assume ‘that the social sciences adhere to a single scientific

method for the justification of their knowledge claim’ (Anderson, 1986, p. 156).

Hirschman (1986, p. 237) believes that the reason of various forms of positivism

dominance in the field has been marketing’s roots in logistics and economic issues (areas

which are preoccupied with the incorporation of concepts like profitability, cost

minimisation and marginal returns).


155

Coinciding however with the broadening of analytical perspectives in the 1980’s, many

prominent names in the field (Belk, 1988; Belk et al., 1989; Hirschman, 1986; Thompson

et al., 1990; Venkatesh, 1992; Firat & Venkatesh, 1995; Thompson & Hirschman, 1995;

Shankar & Patterson, 2001; Goulding, 2005) challenged its defining characteristics and

sought to understand consumer practice in their own language (by adopting interpretivism

approaches to marketing research). This transition has of course not gone unmarked by

defence and reaction (Hunt, 1991), and it is still one that has not been fully embraced

within the discipline.

In addition, a further paradigm of consumer research which is based on a behavioural

conception of consumption flourished around that same period (see also chapter 4). The

behavioural analysis of consumption can well be based on the principles of radical

behaviourism (Foxall, 1986; 1987). It advocates that consumer behaviour is operant,

private events are behavioural in character rather than mental and that the locus of

behavioural control can be found within the environment/situation.

In an attempt to justify and maintain the integrity of such philosophical distinctions,

researchers usually draw on Kuhn’s (1962) incommensurability thesis (as in Davies &

Fitchett, 2005). However, the paradigms above (see table 7.1 and the discussion that

followed the table) and the perspectives that these represent should not be perceived as

prescriptions; rather wide variations and differentiations exist even within the paradigms

themselves. These variations make the boundaries among the paradigms blur and the main

ideas behind them not so clear cut. Hunt (1991) described how ‘the main problem with

realism is that there are so many different kinds of it’ and interpretivism/constructivism

are two approaches that steer a researcher into a particular outlook but even within these

alternatives one can find different research agendas and diverse ways of examining and

understanding the data (Goulding, 1999; 2005). Foxall (2007a; Foxall; 2007b) a supporter


156

of the behavioural approach has also advocated the inclusion of intentional terms when

examining consumers’ behaviour in behavioural terms, thus accomplishing the best of

both approaches (see also chapter 4).

Equally, the common definition of research paradigms based on their connections with

particular research methods (usually quantitative methods for positivism- qualitative for

interpretivism) does not really dictate any actual reasons why the whole range of methods

cannot be used in a way concordant with the different goals and assumptions of the

different paradigms (as in Wilk, 2001; Davies & Fitchett, 2005).

As a foundation statement then, this thesis has been written under two basic beliefs.

Firstly, in a ‘rigorous’ consumer behaviour research all kinds of perspectives are

necessary and all paradigms add up to the maximisation of types of questions researchers

are able to address (Foxall, 1987; Hunt, 1991, p. 41; Foxall, 1995, p. 13; Wilk, 2001).

Secondly, although the question regarding paradigms commensurability has not been

answered yet and paradigms are supposedly different on some or all of their dimensions,

recent voices argue (Szmigin & Foxall, 2000; Cupchic, 2001; Davies & Fitchett, 2005;

Foxall, 2007a; Foxall, 2007b; Foxall, 2008) that in order to re-appraise the dichotomies

and to build bridges between different social ontologies, researchers must engage in a

transcendental act of reflection and look for similarities rather than differences.

7.2.2. Philosophical Position

Both radical and intentional behaviourism are not only distinctive schools of psychology

but also appropriate philosophical frameworks of the science of behaviour. Overall,

radical behaviourism has been connected to a naive, Machian, positivism where no

distinction can be assumed between scientific perception and reality (Foxall, 1995, p. 23).

This study will adopt this positivistic philosophical perspective where reality is perceived


157

as real and apprehensible by researchers. Further, specific philosophical assumptions

underlying this knowledge quest are based on the principles of intentional behaviourism

and are summarised in the following table 7.2.

Table 7.2 Ontology, epistemology and research methods of this study

Ontology

Intentional Behaviourism

Behaviour is the subject matter of psychological inquiry.

The locus of behavioural control is to be found within the

situation.

Discriminative stimuli control operant behaviour.

Intentional terms (in terms of beliefs, attitudes and desires) have

a role in such inquiry because these contribute to explanation.

Intentional terms maintain their linguistic properties but not their

ontological.

Epistemology

Radical Behaviourism

Theory is empirically based and leads to sound interventions to

solve practical problems.

The aim is to establish functional relationships. Relationships

that occur when a change in an independent variable results in a

change in a dependent variable.

Operant behaviour is ‘voluntarily’. It operates upon the

environment to produce consequences which determine its future

occurrence based on a learning history.

Operant behaviour is of two broad kinds: contingency-shaped

that is shaped by direct contact with the environment and rule-

governed that is determined by verbal descriptions of

contingencies.

Research Methods Use of quantitative methods (online survey, informed however

by multiple pilot tests).

Source: Table compiled for this study from the following sources: Foxall, 1995, p. 21–22; Foxall,

2007a; 2007b; 2008; 2010a, p. 50–56.

The specific knowledge inquiry has been reached by adopting a more focused approach to

its research philosophy which is based on the specific pragmatic research needs of this

study. On the grounds of the aforementioned bridging, of what appears to be

incommensurable ideas, the ontology of intentional behaviourism which accommodates

elements of both the intentional and the contextual stance, the epistemology of radical

behaviourism which advocates the importance of functional relations and contingency and


158

rule-governed behaviour and a quantitative research methods approach, are the most

vigorous choice to serve and achieve this study’s objectives.

7.3. Research Design

The following sections will deal with the research design of this study. With an aim to

facilitate understanding of the processes and decisions involved when considering the

design of a research, three aspects of research design are discussed, namely: the research

process (inductive, deductive and abductive reasoning), the research purpose (exploratory,

descriptive and explanatory or causal) and the timeframe of data collection (the difference

between longitudinal and cross-sectional data is of main concern here). A summary of

research objectives and hypotheses in the form of several tables (see tables 7.3; 7.4 and

figure 7.2) describe the ways that the choices of the research design have served the

objectives of this research.

7.3.1. Research Process

Aristotle was the first to summarise two processes of research: inductive and deductive

processes. In the recent history of social sciences many scholars have examined the topic

of these types of reasoning and the implications for research (Blaikie, 2000; Hyde, 2000;

Saunders et al., 2009). It is widely acknowledged that the difference between these

approaches lies with the role and relationship of theory and data collection. The inductive

process moves mainly from the data or observations to viewpoints (theoretical

conceptualisations) while the deductive process moves mainly from viewpoints

(theoretical conceptualisations) to data collection.

The kind of research which emphasises the continuous questioning has been named

abductive reasoning and Rudestam and Newton (2007) prefer to describe this process as a

‘wheel’ to indicate that research is usually a system where parts interact in order to


159

produce the necessary validity (see figure 7.1 below). In abductive reasoning then the

researcher moves constantly between the theory and the findings in order to produce the

necessary understanding.

Advocating the fact that the emphasis of this research was placed on a) the examination of

an existing conceptual model of consumer behaviour (even though the model was

investigated from the different perspectives of an extensional and an intentional

understanding) and equally b) the development of hypotheses based on theoretical

arguments, a deductive research process has been utilised where the formulated

hypotheses could conceivably be falsified by a test on observable data.

7.3.2. Research Purpose

Another useful distinction that should be discussed in this research strategy section is the

distinction among the diverse purposes that a research design can serve. The most

common classifications are exploratory, descriptive and explanatory (or causal) (Kent,

2007; Burns & Bush, 2010). Exploratory research has been widely associated with the

initial stages of a research where topics are examined with no prior knowledge or with a

purpose to shed new light to existing topics (or see familiar topics with new eyes).

Exploratory research is usually used to define terms and generate insights for further

elaboration and additional research. Descriptive research is mainly concerned with the

characteristics of a phenomenon. In marketing it is usually concerned with answering

questions like who, what, where. According to Burns and Bush (2010) it is usually cross-

sectional in nature (meaning that data collection occurs at one point in time and usually

act as a depiction of a population). Descriptive research is finally also suited for

hypothesis testing and can be regarded as an extension, forerunner, originator, or part of

explanatory research (Saunders et al., 2009). That is because, for the explanation of a

phenomenon, an accurate view of its facts, elements and dimensions should be known to


160

facilitate predictions, hypotheses development and examination of relationships.

Explanatory research then attempts to explain the reasons behind situations and

behaviours. When the aim is the determination of cause and effect relationships between

variables these studies are also called causal. The usual way to ascertain causality is the

use of experimental designs where variables are manipulated and effects can be

established.

Although this categorisation of research purposes seems theoretically useful, in practice

according to Kent (2007, p. 12) it is somewhat inadequate to capture the reality of

conducting research. Pure exploratory designs would only involve generating new ideas

and pure descriptive ones would be limited to the analysis of variables one at a time,

however most research projects include an exploratory phase where ideas are refined, a

descriptive phase for description and will then move on to the establishment of possible

relationships between variables (even if these relationships are described as functional and

not causal as is the case in this research).

In an attempt to characterise this research, two of the aforementioned research purposes

will be mainly used in order to adequately answer the problems posed by this research.

The first stage of research instrument development will be described as purely exploratory

because in accordance with the definition of exploratory research this will seek to

generate new insights, critically debate the existing knowledge and elaborate on the

constructs of interest and their proposed conceptualisations. As there are no previous

conceptualisations of contextual confusion, an exploratory pilot test has assisted to move

confidently to subsequent pilot tests and finally create the main data collection instrument.

Subsequently a research approach which carries elements of both exploratory and

descriptive research was undertaken. This kind of research is suitable when testing general

hypotheses and when causality is not a prerequisite of a study.


161

As usual, there are criticisms and shortcomings of the aforementioned research purposes.

Exploratory research is often described as ‘a brief, fleeting, preliminary stage that gives

way to the real thing’ (Stebbins, 2001). However, as described by the same author, the

heart of research in social sciences is exploratory as research in its own right should be

characterised as a learning and cumulative process which is governed by exploratory

choices and personal interest (Stebbins, 2001). In this way no single study can be

definitive and all research is characterised by exploration.

Additionally, although some researchers argue that especially descriptive research is

mainly a technical matter downgrading in that manner the need for theory and agreeing

that merely measuring variables and identifying the relationships and correlations is

enough, Blaikie (2000) counter-argues that all kind of concepts carry theoretical

implications and thus there is no escape or other way to conduct a research but to accept

that there are theoretical implications in any study.

A summary of the way these two research designs are combined in order to serve the

research objectives and questions of each research stage are presented in table 7.3.


162

Table 7.3 Summary of research objectives and questions of each research

stage

Research

stage Objectives Research Questions/ Hypotheses

Exploratory

To assist the process of

questionnaire development by

identifying the conceptualisation

of consumer confusion that best

describes the concept as a state

directed towards specific markets.

This conceptualisation should be

free (as far as possible) from other

states like frustration, bother,

annoyance etc.

What is consumer confusion?

Which of the previous

conceptualisations of confusion (if

any) better depicts the state in

specific markets?

Descriptive

BPM-I. To examine whether

situational contingencies can be

modified by the person’s rule-

making (personal level of

explanation).

Thus, the aim is to identify the

effect of confusion as a punishment

to the situational reinforcements

(utilitarian and informational) and

to behavioural responses.

Hypotheses 1-6

H4: Overall, the range of

confused consumers will indicate

lower levels of Pleasure than the


H5a: Overall, the range of


lower levels of Arousal than the


H5b: Overall, the range of


the same levels of Arousal with

the range of non-confused

consumers.



lower levels of Dominance than

the range of non-confused

consumers.



lower levels of Approach

behaviour than the range of non-

confused consumers.



higher levels of Avoidance

behaviour than the range of non-

confused consumers.


163



lower levels of Aminusa

(Approach-Avoidance) behaviour

than the range of non-confused

consumers.

To identify areas that the

intentional and the contextual

stance are found to interact to

produce an effect on behaviour.

Hypotheses 7-12.

Proposition: The effect of

confusion on reinforcement and

behaviour will depend on the

situational effect of the market.

H7: The effect of confusion on

Pleasure will be stronger for the

market characterised by overall

lower levels of experience.


Arousal will be stronger for the




Dominance will be stronger for

the market characterised by



Approach behaviour will be

stronger for the market

characterised by overall lower

levels of experience.


Avoidance will be stronger for the




Aminusa (approach-avoidance)

will be stronger for the market

characterised by overall lower

levels of experience.

H13: The two markets are

expected to differ in terms of

utilitarian reinforcement with the

high technology market expected

to have higher Pleasure than the

grocery market.


164

BPM-E. To examine the predictive

power of the BPM in terms of the

reinforcing qualities and behaviours

of consumer situations, which are

beyond the original model.

Hypotheses 13-19.



informational reinforcement with

the high technology market

expected to have higher Arousal

than the grocery market.



dominance with the high

technology market expected to

have lower Dominance than the

grocery market.

H16: Possible differences are

expected in the levels of

Confusion between the two

markets that act as discriminative

stimuli in this study.

H17: Approach will be higher in

the market characterised by higher

levels of utilitarian and

informational reinforcement (thus

the high technology purchasing

situation).

H18: Avoidance will be higher in

the market characterised by lower

levels of utilitarian and

information reinforcement (thus

the grocery market is expected to

have higher avoidance).

H19: Aminusa, the net difference

between approach and avoidance

will be higher in the market

characterised by higher levels of

utilitarian and informational

reinforcement (thus the high

technology purchasing situation).

H20: Affective variables of

Pleasure, Arousal and Dominance

will each have a positive

relationship with Approach.

Confusion will have a negative

relationship.


165

To examine whether reinforcement,

aversive consequences and

behaviour setting scope will

determine behavioural variables as

expected by the models.

Hypotheses 20-24.



will each have a negative

relationship with Avoidance.

Confusion will have a positive

relationship.



will each have a positive

relationship with aminusa, the net

difference between approach and

avoidance. Confusion will have a

negative relationship.

H23: Aminusa (the net difference

between approach- avoidance)

will be determined by the

variables Pleasure, Arousal,

Dominance and Confusion.

H24: Two-way interactions can

be identified between the

affective variables Pleasure-

Arousal when examining their

effect on Aminusa.

Source: this study

7.3.3. Data Collection Timeframe

Another important element of any research, positioned within the boundaries of a research

design (Saunders et al., 2009), is the consideration of the element of time in the study of

the topic of interest. This aspect is crucial because it can determine the results and

generalisability expected from a study. Among the diverse research designs proposed

(Bryman & Bell, 2007, p. 44) the cross-sectional and longitudinal approaches are the ones

more closely describing the time frame of data collection. For longitudinal research, time

is a dimension that plays a particular role in addressing the research objectives as it

advocates for data that have been collected in at least two waves on the same variables

and on the same participants. Such data facilitate research objectives connected with

change and evolvement of phenomena (Bryman & Bell, 2007, p. 61). On the other hand,


166

cross-sectional designs study phenomena at one particular point in time and although there

are certainly practical limitations and considerations for this choice, there might equally

be theoretical reasons that support this option according to the study’s objectives. Apart

from the element of time, cross-sectional designs are usually characterised by survey

strategies and the collection of quantitative data (Saunders et al., 2009).

Turning the interest to this study, the approach to data collection adopted is cross-

sectional in nature, where a subset of the population is measured at one specific point in

time. The reason for that choice is twofold; the orientation of this study is beyond

measuring change or development over time, which makes longitudinal data unnecessary.

Cross sectional data are particularly suited and sufficient for the comparison of differences

among situations or groups of participants and establishment of relationships, which is the

case and objective in this inquiry.

7.4. Research Methods

Following reflection on the strengths and limitations of several research methods and

more specifically examining the differences between qualitative and quantitative

approaches (see Appendix 3 for the strengths/weaknesses and overall framework that has

been proposed as more suitable for using each approach based on Mack et al, 20058) and

further the state of the theoretical model of this study, this research has advocated for the

use of a quantitative approach. The development of the research instrument

(questionnaire) has been informed by several meticulous exploratory research studies/

pilot tests (DeVaus, 2001). This meticulous research approach can be justified when

considering the online nature of main data collection. The emphasis has been placed on a

8 The framework proposed and described in Appendix 3 is only suggestive and generalised. It is

acknowledged that there are specific approaches in each method that do not adhere to these rules.


167

quantitative survey; however each step of the process has added the necessary knowledge

for the completion of this project. Figure 7.2 summarises the research steps used in this

study in order to achieve the questionnaire development by adopting an approach based

on 3 research phases.

Figure 7.1 Summary of the research steps for the development of the main research

instrument and the implementation of this study.

Source: this study

This research has been informed by secondary information following a critical evaluation

of the literature (research phase 1). The information gathered helped with the progression

and enrichment of the subsequent stages of collecting data from primary sources (research

phases 2 and 3). However, as the dotted line starting from the literature review and

extending to all other research stages indicates the evaluation of the literature extended

throughout this research project and has enriched all subsequent research stages.

Phase 1

Elaborate market and literature evaluation

Retail shopping

Confusion- the nature of confusion

Literature on emotions

Behaviour Analysis/ BPM/ Intentional Behaviourism

Phase 2

1) Exploratory test: 7 lay participants completed the questionnaire (both

existing confusion scales) and commented / discussed several issues. Some themes and the way forward was exposed from this process.

2) Pilot 1: 6 final confusion scale judges +PAD and Approach-

Avoidance (2 PhD students, 2 English (linguistic) teachers, 2 lay consumers).

3) Pilot 2: 4 academics (3 of them are experts on

the specific topic) completed the questionnaire to further ensure face validity.

4) Pilot 3: Testing the questionnaire with a convenience sample (n=56) through

the staff of two London-based Universities.

Phase 3

Final large scale survey (n=260) through an online research panel.


168

7.4.1. Research Phase 1- Critical Evaluation of the Literature

The collection and critical understanding of literature has been conducted throughout the

duration of this research project. Various types of resources including but not limited to

academic journals, books, past theses, conference proceedings, knowledge over the

internet and workshop and seminars material have been used. Most of these resources

have been collected through scientific databases like: ABI Inform Global (ProQuest),

EBSCO, Emerald Library, Google Scholar, IngentaConnect, PsycINFO, Management and

Business Studies Portal (British Library), ProQuest Dissertation and Theses, Science

Direct, Social Science Citation Index, Scopus and Zetoc. Several keywords organised

through excel files have been used for the more systematic implementation of the

literature task. Keywords used include: retail environment, retail shopping, (consumer)

confusion, emotions, intentionality, behaviourism, BPM and intentional behaviourism.

The Boolean logic (AND, OR, NOT,*, ?) that allows the limitation, expansion or

combination of items found was used (Saunders et al., 2009, p. 83). The logic of

‘snowballing’ in the sense that the reference list of several key articles informed the

choice of further literature search was also implemented.

7.4.2. Research Phase 2 and 3- Use of Different Kinds of Interviews

In terms of the research methods for the collection of primary data an approach based on a

quantitative methodology was used in this study. Although Creswell and Plano Clack

(2007) indicate the diverse ways9 that qualitative and quantitative methods can be used

during a research project, the use of a behavioural approach and especially of the BPM in

this study (a conceptual model that has been developed through previous research and can

offer researchers with firm theoretical guidelines) dictates and justifies the use of a

9 The ways they suggest are: merging of the data, connect the data in a kind of sequential order and finally,

embed the data


169

quantitative approach. The process of preparing and developing the research questionnaire

itself has been however informed by a number of different data, in the form of a small

scale qualitative/ exploratory test and several interactions with knowledgeable and lay

participants (pilot tests). This technique is one frequently utilised due to the nature of

qualitative and quantitative data. Pilot tests and qualitative kind of interactions with

participants is very much suited to generate a deeper understanding of the phenomena in

question and help develop and test the research instrument and quantitative methods can

be used to test specific hypotheses (Patton, 2002).

Table 7.4 presents each of the preliminary methods used, along with goals achieved,

details on the pool of relevant participants, sample design, ways of analysing data and

time periods of data collection.


170

Table 7.4 Details of research methods used in this study for the purposes of

questionnaire development, testing and data collection.

Research

method Goal Pool of participants Sample Analysis

Time

Period

Structured

Personal

Interviews

To facilitate the

development of

the main

quantitative

instrument.

To develop an

understanding of

confusion and

place it in

context.

Cardiff Business

School staff.

Snowballing from

personal

acquaintances.

Non-probability

convenience

sampling.

Continuous

memo-writing,

interpretive and

reflexive

readings,

comparisons of

answers, and

formulation of

the common

understanding,

which was

enriched with

judgement

based on

theoretical

knowledge.

February-

March

2012

Self-

completion of

initial

questionnaires.

To facilitate the

development of

the main

quantitative

instrument by

ensuring a) face

validity and b)

ease of use and

understanding of

the general

structure of the

research

instrument.

Cardiff Business

School staff.

Cardiff University

Invigilators.

Academics in the

field of marketing

and consumer

behaviour.

Non-probability

purposive

sampling.

Comparison of

comments.

Personal

judgement

based on the

requirements of

this research.

March-

June 2012

Questionnaires:

Online and

self-

administered

To test the

research

hypotheses.

Study 1: University

staff.

Study 2: Qualtrics

online research

panel.

Study 1: Non-

probability

convenience

sampling.

Study 2:

Probability

sampling.

EFA,

correlation,

ANOVA,

factorial

ANOVA and

multiple

regression.

Study 1:

July 2012

Study 2:

November

2012

Source: this study

7.4.3. Interviews

Following the description of the way this research employed several steps to questionnaire

development, it is essential to state the different forms of personal interviews as the main

secondary data collection method. Figure 7.4 categorises the diverse research methods

available to researchers based on two dimensions, according to the number of participants


171

involved and the involvement of the researcher in each process. The research methods

repertoire includes different kinds of interviews (ranging from structured to unstructured)

and also involves more unobtrusive methods, like participant observation.

Figure 7.2 Data collection methods

Source: McNeill & Chapman, 2005, p. 22 (original source: Worsley, 1977).

In view of the varied limitations and strengths of these research methods, their practical

implications along with the specific research objectives of this study, both semi-structured

interviews (a combination of open and closed questions used during the small scale

exploratory test and the pilot tests) and a structured questionnaire/ survey (simple,

specific, closed questions) (Gillham, 2000) were utilised in the different stages of this

research. Although social science researchers and generally our society has been called an

‘interview society’, accused of placing a distinctive emphasis on interviewing as a form of

data collection (Atkinson & Silverman, 1997), interviews have also been described as the

‘bastion’ of research in social sciences and their importance is unambiguous (Briggs,

1986).

Nu

mb

ers

invo

lve

d

Personal involvement of

researcher

HighLow

Few

Many

Participant Observation

Surveys

Structured interviews

Semi-structured interviews

Unstructured interviews

Observation


172

The use of interviews is justified in this endeavour considering two aspects of this

research a) the emphasis that radical behaviourism places on verbal behaviour of

individuals (BPM-E) and also b) the necessity to use intentional terms which describe

individuals as bearers of beliefs, desires and propositional attitudes (BPM-I). Such entities

are more precisely described by those bearing them and could not be as accurately

examined by observation or other unobtrusive methods (Murphy & Dingwall, 2003, p.

78). As admitted by Murphy and Dingwall (2003, p. 78) everyday interactions and

discourse might produce such spontaneous narratives of factors that influence behaviour

but there is no predictable place or time where consumers routinely do so. It would be

necessary thus to observe them for an inordinately long period to secure very little data.

The following sections will proceed with the ways that the use of interviews has found

application in the initial exploratory step, the pilot tests and will describe the ways these

research steps have facilitated the development of the main research instrument of this

project.


173

7.4.4. Exploratory (Pilot) Test

a. Semi-Structured Interviews

Aim- sample-criteria for participation

The aim of this exploratory step has been to inform the questionnaire development by

identifying which of the two previous conceptualisations of consumer confusion can find

a better application in a contextual manner. The two confusion scales (see chapter 3, p.

49-52) were previously generated to serve different purposes. However, both

conceptualisations could potentially find application to the measurement of confusion on

different occasions as admitted by the researchers who developed them (Walsh et al.,

2007; Walsh & Mitchell, 2010; Schweizer et al., 2006). Specifically, Walsh et al., (2007)

invite future inquires which will act to extend and test the appropriateness of their scale

for use in specific contexts and shopping situations.

The decision to re-examine an existing scale in different situations and in that manner

broaden its context of applicability, was also based on the acknowledgment of the

increasing voices in the field of advertisement, consumer and marketing research

indicating that theory development and refinement have suffered from the phenomenon of

the ‘isolated study’ indicating the lack of an explicit replication tradition in these fields

(Easley et al., 2000). Although this is not a replication study per se it will seek to examine

which of the previous conceptualisations of confusion best depicts the conception of

confusion as applied to specific shopping environments/markets. By extending the

applicability of this scale to a contextual manner and re-testing its validity/ reliability, the

accretion of knowledge is facilitated.

It was then a conscious decision not to develop a new confusion scale. Two characteristics

were judged as essential for the adoption of one of the two existing conceptualisations of

confusion for this project:


174

1) The conceptualisation should be received with consensus concerning its capacity to

represent confusion in both of the situations in question, and

2) This conceptualisation is not characterised or is minimally characterised by other states

like annoyance and frustration.

In order to implement the aim of this exploratory step a non-probability convenience

sample was judged as sufficient. Seven participants were chosen based on their

willingness to take part and their ability to answer questions regarding confusion. The lack

of specific sample size rules to follow when conducting such essential ‘qualitative’ parts

complicates the choices made at that level. As the main objective of this step has been to

enhance the researcher’s understanding concerning consumers’ thinking and perception of

confusion and connect these ideas with theoretical constructs, interviews stopped when it

was clearly felt that data were enough to inform the study and its results. This approach

has been called theoretical saturation (as in grounded theory) however in this study

judgement saturation seems more appropriate (the point where the researcher felt a choice

could be made with confidence).

Table 7.5 describes the characteristics of this sample:

Table 7.5 Participants of the exploratory test

Gender Age Group

Participant 1 Female 55-64 Vocational/technical school

Participant 2 Female 55-64 High School

Participant 3 Female 65 and over High School

Participant 4 Female 25-34 Graduate School (MSc/PhD)

Participant 5 Female 25-34 Undergraduate Degree (BA/BSc)

Participant 6 Female 45-54 Undergraduate Degree (BA/BSc)

Participant 7 Male 25-34 Graduate School (MSc/PhD)

Source: this study

The over-representation of female participants is not considered problematic, as the

research objectives were of exploratory nature and following the completion of these

seven interviews objectives were evaluated as successfully achieved. The researcher


175

compared and combined the participants’ discourse with previous research, theory and

understanding and a firm decision on the best course of action for the quantitative

measurement of confusion was reached.

One issue identified during the conduct of the interviews concerned the most recent

shopping trip for the majority of these participants. This in most cases was reported to be

a grocery trip, thus consumers in the short first part of the interviews very much focused

on this shopping occasion (this is however probably not connected to the nature of the

sample, more possibly it can be attributed to the frequent and mundane nature of grocery

shopping). This issue was overcome with the introduction of the PC/laptop context in the

next stage of the questionnaires’ administration.

Procedures of the interviews

During these interviews a short interview guide along with four structured questionnaires

(the role of the questionnaires was to act as stimuli for easier elaboration on the topic)

were used in order to achieve the research objectives. To start with, only two initial open

questions were introduced to participants. The first question concerned participants’

definition/understanding of consumer confusion in general and they were then asked if

they could describe a confusing experience they experienced during their most recent

shopping occasion or one that they could remember. This process was very general but

was necessary in order to introduce participants to the subject and empower them to think

in terms of their own perception of the topic.

Subsequently, the two confusion scales were given to the participants in paper (4

questionnaires in total- one of the scales in the two markets and again the other scale in

the other two markets). This part was left more open with no specific questions.

Respondents were left to read (and if they wish complete) the two scales in the two


176

contexts and were then asked to express their own thoughts on the suitability of each in

those two contexts.

Interviews were voice recorded (mainly concerning the first interview part) and then only

partially transcribed. As consumers spend much of the interview time completing and

commenting on the questionnaires, full transcription of the interviews was judged as

unnecessary. Note-taking at the moment of the interview seemed as a meaningful

approach for this part of the research. The analysis was based on multiple and reflective

readings, comparisons of the transcripts and memo-writing following each interview, in

accordance with the principles introduced by Charmaz (2006, p. 80-81). This process

helped to retain the overview of the interviews, to maintain understanding between the

interviews and thus to better appreciate consumers’ discourse and reach conclusions

easier. The findings were further compared with and very much based on researcher’s

judgement and previous theoretical conclusions from relevant research.

Findings

Table 7.6 contains the spontaneous answers on what constitutes consumer confusion

according to the participants:

Table 7.6 Spontaneous answers on the topic of 'consumer confusion'

according to the participants

Abstracts Code given by the

researcher

Participant 1

‘I would say that confusion is like if you have two

sugars for instance...different kinds of sugar which

one is the right one for you to buy? You

know...and actually like that sugar...the half spoon

sugar...that half spoon sugar....I did not realise it

was half sugar and half something else, it must be

the artificial sugar or something in that

packet...and when I did buy it was very powdery

and I did not realise why until somebody else told

me...so that would be confusion.’

Overload- Ambiguity


177

Participant 2

‘Well the best example of confusion that I can

think...is possibly conflicting medical sort of

theories about what is good for you and what is

bad for you that seem to change on a regular

basis...one minute everyone must eat say...berries

and somebody else saying no they are actually bad

for you.....’

Ambiguity

Participant 3

‘Sometimes it is packaging, especially when they

re-package something you go and you know you

could have bought that package in grey with black

writing and they change the packaging and you

are looking at it and then you have to start and

read it to make sure it is the one that you bought

before...and it is still suitable for vegetarians

and’...

Novelty

Participant 4

‘Confusion can happen with the pricing because

particularly when there are offers, at least at the

super market they write the price of products by

weight or by volume so you can compare but

sometimes especially in other kind of stores like

when buying a computer, like here for example,

you need to compare all kind of information and

that might also be confusing if you need to

arrange your budget as well...sometimes the

selection it is not quite...if there is so much

selection it might take more time to

choose...Sometimes you cannot really feel the

difference...if it is just small differences...is it

really worth buying the most expensive? ...’

Ambiguous pricing

information,

Overload of products and

offers

Participant 5

‘and I would define as confusing the case when

you know what you are looking for but cannot

really locate it in a store or even in a website’.

AND

‘I suppose...I suppose the advertising of all the

products is misleading ehm...for instance there is a

product and I cannot even remember what it is but

its beauty type product and I think at the bottom of

the ad ...I think they even said that it wasn’t used

in the ad...the disclaimers that they actually put at

the bottom of these ads...you know...your eyelashes

are going to be like this long meanwhile it says

false eyelashes used- you know- I just find this

misleading and confusing’

Novelty

Ambiguity

Participant 6

‘that is for me the biggest confusion when I go to

any kind of store, if the price is not displayed

correctly or if it is not there at all. And also

when...you know if something is in promotion and

it’s not, they should...for me that is absolutely

appalling because if in promotion obviously you

got to go quite fast, if it’s not there I am not

buying anything you are losing business, I don’t

get it...’

Ambiguity


178

Participant 7

‘I think...confusing...let me think...sometimes there

are just too many products or in terms of

information, yes...finding the one for you.... Even

when you know the brands and the characteristics

to look for but yes I still think...when there is...how

do you call it, too much of everything’

Overload

Source: this study

A general but important observation regarding the defining characteristic of the state of

confusion is that most consumers add the element of ‘self-related consequences’ in their

answers. ‘Not being able to do the best for me’ concerning products, prices, nutritional or

general information and due to several market reasons- ambiguous information, overload

etc., seems to characterise confusion beyond all else.

Adopting the reflective approach characterising this data analysis, the interviews indicated

that the ‘overload-similarity and ambiguity’ scale more accurately characterise the state of

confusion in diverse settings. The conceptualisation introduced by this scale is more

general and can find application in more shopping situations including online shopping.

Although the similarity factor was not spontaneously mentioned by any of the

respondents, consensus was achieved that similarity of products and brand logos are

indeed a source of confusion. In addition, overload and ambiguity were unambiguously

mentioned (even when no specific stimuli were provided) by the majority of individuals

as representing confusion (as in table 7.6).

The environmental scale at the other end was suited in a retail environment like a grocery

store but could not accurately capture the reality in other situations. Participants found it

difficult for example to answer several items when the scale was applied in a PC/Laptop

situation. Scale items on the loyalty card data (stimuli conflict) or questions regarding the

product packaging which is not such an important and obvious element in a PC/laptop

setting, added to participants’ frustration and uncertainty when completing this scale.


179

The most important setback with this measurement scale was however that some of its

main elements were not described as confusing but rather as frustrating or annoying by a

majority of participants. Not all participants found elements like crowding and long

waiting queues (comfort factor) or continuously changing the position of items and

products been replaced too often in a retail setting (novelty factor) as confusing e.g.:

‘once I was looking for coconut and I looked and I looked and I was getting quite annoyed

that I could not find it because then you have to go to another shop to find it but actually I

did find it at the end but I would not think of that as confusing, really frustrating yes....’.

(participant 1)

‘the worst thing is when they move everything round or the store is really scrappy and you

go in any store from jeans and jumpers to grocery stores and you are in a hurry...and you

go down to where you know the product and everything is and it isn’t there any

more...and oh...no it’s not there or we have moved it down to the end of the

store...why?????? this is annoying but I wouldn’t say confusing’. (participant 3)

‘but I don’t regard that (not being able to locate products because of moving or replacing

them) as confusing because every sort of store has their own way of displaying their

products in different areas...I don’t know’. (participant 6)

‘confusion...no it doesn’t ...I don’t really go down that road like when products are

missing from shelves and you are in a hurry and you cannot really find what you are

looking for...’(participant 7)

‘This is actually a bad experience when...you know having all these people in the aisles

and you cannot move, you cannot stand- you know- on Saturday and Sunday...like

weekends shopping and I am thinking...Oh my God! I can’t endure it. That for me would

be an absolute nightmare- absolute nightmare.’ (participant 6)

Considering the earlier demarcation of the terms ‘confusion’ and ‘frustration’ in

psychological research (see chapter 3) this finding does not come as an imposing finding.

Confusion is characterised by a lack of understanding and elements like crowding or

loyalty card data management do not seem to be part of confusion but rather frustration. In

addition, previous consumer behaviour research has never found evidence which could

corroborate the addition of crowding as a confusing element. Crowding has been

described in the past as stressful (Aylott & Mitchell, 1999) and dissatisfying (Machleit et

al., 2000) but never as confusing.


180

To finish with, the latter environmental scale was overall more difficult in administration

and had on average a longer completion time. On all of these grounds, a conceptualisation

based on the ‘overload-similarity and ambiguity’ scale was assessed as an adequate

representation of the concept of confusion in this study that can find unequivocal

application in different settings. It is more representative of the concept of confusion,

minimising the risk of measuring other elements like frustration or annoyance.

7.4.5. Pilot Tests

a. Personal Interviews- Self Completion of Initial Questionnaires (pilot

test 1)

Aim- Sample- Criteria for participation

This second stage of the instrument development has used more structured personal

interviews (10 participants agreed to take part in this stage) in order to serve a very

specific aim: to improve and finalise several elements of the research questionnaire. After

establishing which existing conceptualisation of confusion is more appropriate to be

adopted in this study, the following elements were checked:

Were the questions placed in the best format and order?

Were questions well received and understandable by the respondents?

Were the instructions comprehensive and detailed on what was expected to do?

Was the change of situation (market) clear to the respondents each time?

Although the questionnaire was mainly created based on the previous critical evaluation

of the literature and the exploratory stage, this step was also judged as essential. The

understanding and ease of use of the main elements of the questionnaire are imperative

and should be established in every survey research (De Vaus, 2002). However, due to the

nature of this instrument, a self-paced, self-administered online questionnaire, any sources

of misunderstanding should be eliminated as these could impose a danger to the validity


181

and reliability of the results. In addition, several problems which were identified from the

previous step were brought to the fore in this phase for further elaboration.

As the aim of this stage was to increase the readability of the questionnaire,

knowledgeable participants on aspects of language and questionnaire structure along with

lay people were identified as the best source of information. This stage then took place

between April and June 2012. Five knowledgeable participants who all had relevant

market experience (including 2 PhD candidates of consumer behaviour, 2

English/linguistic teachers and 2 lay people) and 4 academics (3 of which are experts on

the specific topic) agreed to take part. The inclusion criteria for this stage were then

multiple: 1) willingness to take part, 2) good knowledge of the English language and 3)

ability to provide relevant information.

Procedures of the interviews

A copy of the questionnaire was provided to all participants who were asked to fully

complete it based on their judgement. No additional ethical form was provided as the first

page of the questionnaire clearly indicated the relevant ethical issues, which are applied in

the case of this research phase. Participants were then asked to comment on unusual or

confusing issues especially when it comes to its wording, structure and content. Some of

the participants preferred to write down their comments but some chose to examine the

questionnaire question by question and comment separately on each part. Two of the

participants (the linguistic teachers) required more time to examine the questionnaire and

returned it the next day along with their comments. The whole process was then left open

and up to the discretion of each participant. If an issue was of concern to the researcher

and it was not mentioned by one of the participants they were asked to comment on that.

These interviews were not voice recorded. Notes-keeping was the preferred method for


182

this stage. In most cases, questionnaires were returned to the researcher and the key

observations of each participant were written on them.

The analysis has been based on comparisons between participants’ comments and

reflective practices achieved through the negotiation of participants’ comments and

researcher’s knowledge of the topic/theories applied and research objectives.

Findings

Through this process the approximate time of completing the questionnaire was

established to 20-25 minutes depending on the individual. Also several questionnaire

improvements were implemented. The confusion scale had several problems of wording

and phrases that are not widely used or were perceived as obsolete/pretentious. The phrase

‘host of stores’ was replaced by ‘too many stores’, ‘between scores of similar products’

was replaced by ‘among similar products’, the verb ‘detect’ was replaced by ‘spot’,

‘owing to the great similarity of many products’ was assessed as being too pretentious and

was replaced by ‘there are many similar products’. Another important contribution of this

stage has been the addition of the words groceries or PC/laptops in the confusion scale.

The original scale was generally referring to products and information and it was initially

decided to leave the scale intact. The participants would then use the questionnaire

instructions to understand which market the scale is referring to at each stage. Most

participants found this unclear and all advocated for the inclusion of the specific words in

the final scales so that these better represent each of the two market situations. For

example, the statement ‘there are many similar products’ was replaced by ‘there are

many similar grocery (or PC/laptop) products’.

Some other comments regarding the emotional/PAD scale were brought to the fore. Both

of the linguistic teachers and one academic suggested for example that ‘relaxed’ is not an


183

exact opposite of ‘bored’ (as in the pleasure dimension of PAD). Such suggestions were

not implemented as the PAD scale has been widely used in previous research. During this

phase, it was evaluated as essential to perform some changes to one of the behavioural

items of the variable Approach, ‘how much time do you spend in a store?’ This decision

was based on its inadequacy to capture the reality of people in retail settings. Mehrabian

and Russell (1974) advocate the adjustment of the behavioural scales based on specific

research situations and this critical step assured that the item would fit the research

context. The general presentation and instructions of the questionnaire were appraised as

adequate with only a few format changes necessary.

b. Online Self Completion of Questionnaires (pilot test 2)

Aim- Sample- Criteria for participation

This second phase of pilot testing took the form of participants answering the

questionnaire online. An invitation to answer the questionnaire was sent to some staff of

two London-based universities. Out of 79 clicks on the survey link, 56 completed

questionnaires were received. Sample characteristics for this pilot test are presented in

Appendix 4. Out of these 56 participants, only 6 buy groceries online and the rest in-store,

while 33 buy technological products online and 23 in-store.

Findings

The aspects examined through these 56 completed questionnaires were in accordance with

De Vaus, (2002, p. 116). The testing included examination of the levels of missing data

and non-response for specific questions, relative variation in answers in order to ensure

that questions will be useful in subsequent analysis; finally, aspects of meaning

misunderstanding and any other difficulties or errors were asked to be raised by the

participants by contacting the researcher through email. Participants of this pilot test were


184

university staff and thus knowledgeable in the conduct of research. Being asked to contact

the researcher in case of being encountered with difficulties in completing the

questionnaire was not seen as incongruent. One email was received by one participant

who indicated two unintended spelling mistakes in the questionnaire.

Further to that at the level of analysis, not many results could be drawn from only 56

participants, however responses were coded and a general rehearsal on the way analysis

would be conducted was implemented. The way that some survey mechanisms, like the

‘block randomisation option’, work was also tested (see section 7.4.6 on the questionnaire

design for more details on this option). The mean time of answering the questionnaire was

17 minutes.

7.4.6. The Research Instrument- Questionnaire

a. Designing the Questionnaire

Several decisions were taken regarding the format of the questionnaire (the final research

questionnaire is attached in Appendix 7). To start with verbal descriptions of the situations

was the preferred method in order to describe the situations. Foxall (1997c) describes that

the use of verbal description of the situations is a better research approach when compared

with the use of more sophisticated techniques like videos, slides or photographs of the

same situations. In the case of written descriptions consumers reflect on the situations and

make use of their own learning history and previous experiences in similar conditions.

The use of photos, slides or videos could evoke specific reactions to the depicted scenes

rather than compel consumers to use their own rules and previous learning in the specific

settings.

Regarding the presentation and description of market situations, an approach based on

simplicity was preferred. The situations were presented as simply as possible, trying to

accomplish the integration of the respondent with the situation. Participants were asked to


185

imagine they are in the specific situations and then answer the relevant questions. The

most challenging issue was whether to include the online element in the descriptions of

the market situations. Online shopping has become a prevalent medium of shopping in

many markets. It was however decided to include the description of online shopping only

in the high/ technology market. Despite the impressive growth (134% growth in value

over the period 2005–2009) of the UK online grocery retailing, it still accounts for only

3% of sales of the total grocery sector, making it a niche channel when estimated in the

broader context (Mintel, 2009). Findings from this study (see chapter 8 on analysis)

supported this decision.

Additional issues of sections and questions’ order were also considered. The questionnaire

commenced with some guidance on the topic and the relevant informed consent (see

ethics section below). The questionnaire comprised of two main sections. Section A

concerned consumers’ socio-demographic information. Specifically, information on age,

gender, higher completed educational level, ethnic group, household size, working status

and finally, some buying habits were included in this section.

Although it is a widely adopted practice to include the socio-demographic information of

the consumers at the end of the questionnaire (Parasuraman et al., 2004), this study has

preferred to ask this information at the beginning of the questionnaire. The reason behind

this choice is based on the following reasons: First of all, it was seen as imperative not to

lose valuable socio-demographic information from respondents. Secondly, it was

appraised that the level of information asked was not going to cause any harm or

discomfort to participants and thus it was seen as harmless for the response rate of this

research to include it at the beginning; thirdly, this information acted as a pattern that

facilitated the collection of data based on approximate sample quota. For all these reasons

socio-demographic information was seen as essential to be presented at the beginning.


186

Following the collection of sociodemographic information, the subsequent section,

section B, has been the main part of the questionnaire were the description of situations

was given to consumers. The emotional variables of PAD were presented before all other

variables for each market, followed by the behavioural approach-avoidance variables and

the multi-item confusion scale was presented last. This way of presentation was decided

based on the consideration that the answers to the emotional and behavioural variables

should be left unbiased from participants’ answers to the confusion scale. The experience

items along with the questions of frequency of shopping followed (see Appendix 7 for the

questionnaire).

It is also essential to mention that the presentation of the markets was randomly assigned

using the ‘block randomisation option’ of Qualtrics software. In this instance,

approximately equal numbers of participants answered the grocery market first and

PC/Laptop second (134 participants) and 126 participants answered the PC/laptop market

first and the grocery next. The choice was unsystematic and was based on a random

procedure chosen by the program. This process was chosen in order to minimise any order

bias that could be imposed by the order of market presentation. If one of the two markets

was consistently presented second, this could invalidate the results and question the

validity of the findings as the constant presentation of one of the markets first could cause

participants to understand the objectives and lead their answers to the ones they thought as

desirables. This problem was easy to overcome by the process of random presentation of

the markets introduced.

b. Measures

Mehrabian and Russell‘s (1974) scales of the measurement of affective responses of

Pleasure, Arousal, and Dominance (PAD) were used without modification in this study.

These three variables are constructed after the semantic differential approach (Foxall &


187

Greenley, 1998; Soriano et al., 2002). Each affective variable was measured on six items

in terms of which the situation in question was rated on a nine-point scale. The Pleasure

(e.g., unsatisfied – satisfied) dimension ranged from extreme feelings of dissatisfaction to

extreme satisfaction. The Arousal (e.g., calm – excited) dimensions ranged from extreme

feelings of calmness to extreme excitement. The Dominance (e.g., submission –

dominance) dimension ranged from extreme feelings of lack of control upon one‘s

environment to feelings of being extremely in control. Following the original instructions

by Mehrabian and Russell (1974) three of each PAD six items (nine in total) were

inverted in their direction in order to minimise bias and all the items were presented in a

random order.

Approach–avoidance was measured by means of six of Mehrabian and Russell’s (1974)

eight statements for these items (those to do with thinking out a difficult task and working

in the situation have been described as inappropriate to consumer behaviour). The six

statements selected in accordance with all previous research of the BPM were, for

approach, ‘How much time would you like to spend in this situation?’, ‘Once in this

situation, how much would you enjoy exploring around?’ and ‘To what extent is this a

situation in which you would feel friendly and talkative to a stranger who happens to be

near you?’ and for avoidance, ‘How much would you try to leave or get out of this

situation?’, ‘How much would you try to avoid any looking around or exploration in this

situation?’ and ‘Is this a situation in which you might try to avoid other people, avoid

having to talk to them?’. Confusion was measured by the Walsh & Mitchell (2007; 2010)

scale and market experience with one item measuring the overall market experience with

a market.

Finally, general market experience was measured with a single item. As described by

Walsh and Mitchell (2010) single item use is increasingly common in marketing research


188

(see also van Birgelen et al., 2001). The most substantive issue in this occasion is whether

the item is sufficient to measure the construct. In this case what the study is examining is

not specific experience with aspects of the market/product design, use or price, but is

more concerned with the (general) macro experience, which is a more overall evaluation.

In that sense a multi-item scale is not needed to capture the nature of this macro-concept.

This is in-line with Bergvist and Rossiter (2007) who explain that for constructs that

consist of a singular object, single-item measures should be preferred. Fuchs and

Diamantopoulos (2009) provide a detailed account of the rationale for and against multi

and single item scales. They explain that single item constructs should be preferred when

the construct does not act as a dependent or independent variable but rather as a control

variable and when very high precision is not an absolute necessity as in this case. Also,

multiple-item measurement instruments can occasionally aggravate respondent behaviour,

overload respondents and undermine their reliability (Drolet & Morrison, 2001; Fuchs &

Diamantopoulos, 2009), supporting the use of a single-item scale in the case of market

experience measurement. On these grounds, it was decided that in order to capture market

experience one item would be sufficient for the purposes of this study. Item used was

adapted from Laroche et al., 2003.

7.4.7. Sample Procedure and Size

A study’s population consists of the entire body of units of interest to decision makers in a

situation and a sample is a subset of this population. Samples are often described in

representative terms, but this is not always feasible or even desirable (Parasuraman et al.,

2004, p. 356). A sampling frame is a listing of population units from which a sample is

chosen (Parasuraman et al., 2004, p. 356). In this study the population is the general

consumers who have any levels of shopping experience, the sampling frame is the

participants of an online research panel and the sample size and method of choice for this


189

study was selected in light of the diverse techniques used and was also based on the

purposes of this study.

Several sampling methodologies have been proposed which can be classified under two

broad categories: probability and non-probability sampling. Probability sampling is ‘an

objective procedure in which the probability of selection is known in advance for each

population unit’; simple random sampling, stratified sampling and cluster sampling, are

some of the choices of probability sampling available to researchers. At the other end,

non-probability sampling is ‘a subjective procedure is which the probability of selection

of each population unit is unknown beforehand’ (text for probability-non-probability

sampling is from Parasuraman et al., 2004, p. 360-362); methods like convenience

sampling, judgement and quota sampling are examples of this approach. This study used a

simple random sampling with an aim to ensure a varied sample.

Continuing on the topic of sample size, several considerations guide the decision on the

necessary and adequate sample size to perform statistical tests. Specific research and

statistical requirements have been reckoned before deciding on the necessary size for this

study.

Especially the literature on factor analysis is characterised by the topic of the appropriate

sample size principally when it comes to pure factor analytic studies. According to

Comrey and Lee (1992) (as cited in MacCallun et al., 1999) for the purposes of factor

analysis the categorisation of adequate sample sizes is as following: 100 poor, 200 fair,

300 good, 500 very good and 1000 or more excellent. Although such rules of thumb have

been criticised for their inadequacy to capture the requirements of different studies, the

alternatives proposed usually entail knowledge of previous comparable studies, like for

example knowledge of communalities produced by similar scales during previous factor


190

analysis (MacCallun et al., 1999). This is not always feasible. In order to overcome this

problem another rule of thumb guided the choice in this case. Specifically, Tabachnick

and Fidell (2007, p. 613) propose that ‘it is comforting to have at least 300 cases for

factor analysis’. This analysis has more than 500 cases for factor analysis which has been

described as a very good size.

Overall, such sample numbers are the rule of thumb for conducting General Linear Model

(GLM) approaches like ANOVA and regression. Specifically, Tabachnick and Fidell

(2007; see also Pallant, 2010) describe how a sample of 200 cases or individuals is usually

adequate for safely performing most statistics and it is even an adequate sample to assume

that parametric statistical tests would indicate safe results even if the assumption of

normality is somewhat violated. In addition, it has been suggested that when it comes to

multiple regression (Stevens, 1996, cited in Pallant, 2010, p. 148), ‘about 15 subjects per

predictor are needed for a reliable equation‘. Thus the sample of 260 (520 when both

situations are studied together) used in this research is more than adequate to serve the

statistics and purposes of this study.

7.4.8. Data Collection Procedures

This section will highlight the way quantitative data were collected. The method of online

collection of data and specifically an online research panel based in the UK was the

preferred method. The choice of the research panel was taken after weighting the benefits

against the disadvantages of using online panels. The rational for choosing them will be

explained, details of the data collection process will be described and reflections regarding

the experience of cooperating with a research agency which can facilitate other

researchers with their decision-making will be provided.


191

a. Online Research Panels

The wide penetration and technological advancement the internet is offering has increased

the use of online data collection processes. The promise of online data collection mainly

concerns the easier reach of larger samples combined with the reduction in the complexity

of accessing special populations. In order for research companies to respond to this need

for increasing research participants, to overcome participation fatigue and resistance and

as a result facilitate online research, online research panels have been introduced. An

online research panel is a pool of registered persons who have agreed to take part in

online studies on a regular basis (Göritz et al., 2000). The use of online research panels

has increased over the recent years in both academic and market research (Göritz et al.,

2000). In the same manner as any other type of data collection, online panels have

advantages and disadvantages which are summarised in table 7.7.


192

Table 7.7 Advantages and disadvantages of online research panels for

(academic) research

Advantages Disadvantages

Internet access has now reached 80% of

total households in the UK (ONS, 2012).

More research participants can now be

reached.

No internet access for certain populations.

Online research panels are suitable for ‘hard

to reach groups’ and representative or

probability samples.

Rapid and timely data gathering process.

Less non-response and coverage issues.

The collection of longitudinal data is

facilitated.

Less ethical issues as participants in research

panels have already agreed to be contacted

for research purposes.

Less costly compared to organising

telephone or face to face interviews which

require more intense man labour.

No face to face interaction with respondents

means minimisation of social desirability

bias.

No face to face interaction with respondents

means the research instrument should be

pilot tested well before use.

Online data quality is promoted through an

evolving set of guidelines and standards by

industry and professional associations.

Panel companies are actively designing new

programs and procedures aimed at

validating panellists more carefully and

eliminating duplicate members or false

identities from their databases.

More research efforts are directed towards

an understanding of what drives panellist

behaviours and attempt to design techniques

to reduce the impact of those behaviours on

survey results.

Some panellists might complete large

numbers of surveys (what is often called as

‘professional’ respondents) or complete the

same survey multiple times (by being

registered for example with multiple email

accounts or IP addresses in the same

database).

Data collection is conducted in several

phases (waves). Data collection

improvements can take place (see below

how this measure took effect in this project).

Researchers cannot maintain a control over

the process and the data quality. The whole

process very much relies on developing a

trusting relationship with the research panel

firm.

Source: Table compiled through 1) A reflection of this research experience, 2) Baker et al., 2010,

3) Dennis, 2001, 4) Göritz, 2004, 5) Göritz, 2007.

The table indicates that the use of online panels bears all the positive and negative aspects

of conducting online research in general. In the case of this research an online panel was

used in order to secure a diverse sample that could bear the characteristics of real

consumers and the UK population and which can differentiate this sample from the


193

convenient but widespread use of student participants (see Wells, 2001). In addition, a

rapid, more accurate and reliable online data collection process is secured.

As described in table 7.8 one of the central threats to the research integrity when using

research panels is the quality of the panel participants and the fact that over time they can

turn into ‘professional’ respondents. It has been discussed in previous research that many

psychological mechanisms play a role in this occasion (Dennis, 2001). For example

completion of multiple questionnaires can increase response bias by augmenting

participants’ sensitivity to what is expected from their answers each time and change their

attitudes accordingly (Dennis, 2001). Following email communication with the research

panel provider for this research the aforementioned issue was clarified as following. Each

member of the specific panel receive on average three to four email invitations to surveys

per month and are getting paid between £0.5 to £1 depending on 1) how long these

respondents have been registered with the panel and 2) the topic, length and difficulty of

the survey. In addition, survey sampling is controlled in that way so that the same

panellists are not assigned more than one survey on the same or similar topic (in this case

retail shopping) in at least a four-month period. It is essential but self-evident to note the

importance of developing a trusting communication and relationship between the

researcher and the panel provider throughout the duration of the project. For the better

administration of this research a dedicated project manager was assigned. That person

acted as the point of communication with the researcher and as the sole responsible for the

accurate completion of data collection. This practice was helpful and effective in terms of

providing and receiving information from the same source.

Actual data collection was implemented in November 2012. In order to ensure the

optimum data quality, three waves of email invitations were sent by the research agency

to their participants. The first wave of emails was called a ‘soft launch’ phase. The


194

purpose of this test according to the research panel agency was to examine for the level of

missing data, fully completed responses, and further test the time of completion. Such

simple statistics could indicate questionnaire issues that could be resolved before fully

collecting the data. Following the first wave of responses it was spotted that out of a

hundred- 6 questionnaires were answered in less than 6 minutes; this timeframe to answer

the whole questionnaire was evaluated as unfeasible. The average time of completing the

questionnaire was estimated at approximately twenty minutes and anything less than eight

minutes should be received with reservation. These 6 questionnaires were then checked,

set response bias was identified and these were eliminated from the count of completed

questionnaires. It was also decided that for the duration of the data collection, time

limitation was set to six minutes (meaning that participants submitting the questionnaire in

less than six minutes were eliminated from the count of fully completed questionnaires).

The next two data collection waves were almost simultaneous. The vast majority of

completed questionnaires were completed through the second wave and the third wave

only acted to improve and add to the socio-demographic diversity of the sample.

The markets were presented in a random order in order to avoid creating bias and this

resulted to a situation where 134 respondents answered grocery shopping first while 126

answered high technology, PC/Laptop shopping first. This random choice was valued as a

sufficient variation for evading a biased context. A biased context was further avoided by

the integration of other measurements like market experience and frequency of purchase.

7.4.9. Praxiology

Following the description of the philosophical framework underpinning this study, the

research design and finally the methodological approach and considerations, this section

will guide the reader through another important element of research methodology and also

of the philosophical research paradigm, that of praxiology. Based on Habermas (1993) (as


195

in Mingers & Brocklesby, 1997) this will be subdivided into efficiency and research

ethics.

a. Efficiency

Regarding the issue of efficiency, the research questions are tailored to the specific

philosophical paradigm as these have been developed to explore the propositions put

forward by intentional behaviourism (Foxall, 2004; Foxall, 2007b) regarding the

relevance of applying intentional terms in endeavours theoretically guided by the

principles of behaviour analysis. This research has been developed within the Consumer

Behaviour Analysis Research Group and has been received with enthusiasm by all

relevant research members. Several research method seminars, including those for

research methods, statistics, publications and ways of effective time/project management

and writing, all provided by Cardiff University Graduate School and participation in

research festivals and two Doctoral Colloquiums (part of academic conferences10

) have

provided further understanding of the topic and opportunities for discussion and

elaboration.

b. Ethics

The considerations of ethical mechanisms have become central to the design and practice

of social research. Researchers are expected to act in an informed way, with the physical

and psychological wellbeing of participants, and their right to be informed of every

research aspect being the main concern. Codes of ethical practice have been introduced by

the major research societies (e.g. The British Psychological Society, 2009; ESRC

Research Ethics Framework (REF), 2010), while every research project should now have

10 While developing this research topic, the author has participated in the International Marketing Trends

Conference, Venice (January, 2010), the 2nd

Biennial Academy of Marketing Science Doctoral Colloquium,

Reims (July, 2011) and the 5th

ESRC Research Methods Festival, Oxford (July, 2012).


196

the ethical approval of an appropriate ethics committee. Since the introduction of the Data

Protection Act (1998) which came into effect on 1 March 2000, the concern with issues

related to anonymity and privacy is no longer simply a matter of ethical conduct; it can

also have legal implications.

But how does one go about deciding what is the best course of ethical actions a researcher

should take? Should researchers just follow whatever their instinct tells them? Do they

just follow what people in authority (previous research, ethics committees) propose or did

in the past? Do they just go with what the law dictates? The present research followed a

middle ground between the two key approaches to answer such questions. Between the

teleological approach (the practice of estimating the likely outcomes of a given course of

action, and then choosing the method that has the most positive consequences and the

fewest negative consequences) and the deontological practices (a duty-based system that

directly and simply explain what the code of ethical practice is) (Larry & Michael, 2012),

this research considered both the requirements of the ethical committee but also tried to

tailor these to the needs of the specific research.

Within this framework several ethical issues concerning this research have been

considered. Starting with, the research itself is not related to aspects of personal freedom,

vulnerable groups and it does not concern a particularly extreme, sensitive or personal

topic. Even though an examination of confusion might create the phenomenon of ‘self-

affirmation’, this issue is mainly a methodological concern rather than an ethical one.

‘Self-affirmation’ theory (Steele, 1988) has been proposed as an extension of cognitive-

dissonance theory. The theory proposes that the motivation to affirm the integrity of one’s

self image is very strong and can cause embarrassment, discomfort and attitude change. In

this case admitting confusion in the marketplace, a situation that can be perceived as

participants’ inadequacy to perform, could potentially cause discomfort or embarrassment


197

(Steele, 1988). However, such bias was not evident during the initial qualitative- pilot

interviews, where participants freely expressed their views, without constraints.

Quantitative data also showed evidence of being free from self affirmation. This problem

seems then not to be so strong when considering the issue of consumer confusion.

Focusing further on the complex issue of anonymity and confidentiality, Singer et al.,

(1992; Singer et al., 1995) concluded that mentioning the issue of confidentiality in cases

of surveys of not sensitive topics might have an adverse effect on response rates. If the

survey is indeed sensitive, respondents' willingness to participate is likely to be low

originally, but some of the concerns raised may be reduced by assurances of

confidentiality. Supposing, however, that the topic is not sensitive, respondents'

willingness to participate is likely to be high. If the researcher nevertheless introduces an

assurance of confidentiality, this may suggest to respondents that the questions asked are

more sensitive than the topic implies. Hence, assurances of confidentiality may actually

decrease the response rate in this case. No matter the importance of this observation,

issues of confidentiality and anonymity were raised at the informed consent of this study.

On these grounds, an appropriate full ethical approval form was submitted and approved

by the Cardiff Business School Ethics Committee (see Appendix 8). The form detailed

issues of informed consent, anonymity and confidentiality, sampling, recruitment and

secure data storage. There were two issues that made the data collection process more

demanding in terms of ethics than similar designs. First of all, since data were collected

with the help of a third party (research agency) the communication and arrangements

between the researcher and the research agency have been of outmost importance.

Secondly, the use of an internet-based approach requires review of all the basic research

processes in order to avoid any possible negative implications of research online (security

of data transmission and storage, potential exposure of confidential data, absence of the


198

researcher during the research process as in Nosek et al., 2002; Ess & the AoIR

(Association of Internet Researchers) ethics committee, 2002; Eynon, et al., 2008).

As the main data collection method was an online survey, the first two pages of the survey

would act as informed consent and explain consumers the above issues. The first page

then explained that the survey will be part of a PhD research project at Cardiff Business

School. Other issues covered included the aim of the research and the maximum time

required to complete the survey. Participants were further provided names and contact

details of the researcher and primary supervisor in case they either wished to get a copy of

the report’s findings or wished to discuss any concerns they had regarding the research

process.

An issue requiring further detailing is the issue of anonymity. During this research,

personal details were known only to the research agency which posses all necessary

consent forms and extended knowledge on dealing with personal information; such

personal information was not transmitted to the researcher. In addition, providing names

or signing the informed consent was not necessary as participants indicated their consent

electronically. The problem with this practice is that if data cannot be attributed to specific

participants (complete anonymity) the option of deleting data at a later point in time, in

case participants decide to withdraw from the study for any reason, is unavailable in this

study.

7.5. Data Analysis Techniques

7.5.1. Hypothesis Testing

According to Burns (2000, p. 6-7) hypothesis testing is the ‘systematic creation of a

hypothesis and subjecting it to an empirical test’. Essentially, hypotheses (see chapter 6)

are the statements about population parameters like expected value and variance. In each


199

problem considered, the question of interest is simplified into two competing

claims/hypotheses; the null hypothesis (H0) is a hypothesis of no relationship between the

two variables (Bryman & Bell, 2007), against the alternative hypothesis, denoted H1. In

essence, the statistical outcome of a hypothesis test is to ‘reject H0 in favour of H1’ or

‘not to reject H0’ based on sample information. This principle has facilitated the

formation of accurate hypotheses which along with the use of appropriate statistical

techniques have dealt with the considerations of minimising Type I and Type II errors.

7.5.2. Descriptive and Uni-Bi-Multi-Variate Techniques

In this manner, several steps have been adopted in order to accurately analyse the data of

the questionnaires. First a description of the socio-demographic profile of the sample is an

essential step in order to understand the ‘who’ of this survey. The next step is the

preliminary analysis, which includes all necessary elements of data management like

exploration of the data for missing data, outliers and normality and finally provision of

some descriptive statistics for the items. The psychometric properties of the scales were

next examined. Following this preliminary analysis, which lays the foundations for the

more valid and reliable analysis of hypotheses, several univariate, bivariate and

multivariate statistical techniques have been applied to the data in order to test the

hypotheses.

Multivariate analysis and methods are not easy to define. According to Hair et al., (1998,

p. 6) ‘it refers to all statistical methods that simultaneously analyse multiple

measurements on each individual or object under investigation’. In this manner any

simultaneous analysis of more than two variables is considered multivariate while

univariate refers to the analysis of one variable only and bivariate to the simultaneous

analysis of two variables.


200

Factor analysis is the first method used with an aim to examine the unidimensionality of

the scales (the idea that several items are strongly associated and present a single

concept). Factor analysis is a very generic name given to a class of multivariate

techniques which address the problem of analysing the structure of the correlations among

a large number of variables (Hair et al., 1998, p. 90). In this way data reduction can be

achieved by substituting each underlying dimension for their original variable. One further

important but generic issue needs to be addressed in this initial explanation of factor

analysis. There are two approaches that one can take to conduct factor analysis: one is

exploratory and the other confirmatory (Hair et al., 1998; Pallant, 2010). Exploratory

factor analysis (EFA) is often used in order to explore the possible interrelationships

between the variables without imposing a structure to the outcome based on theoretical

support or prior research. Confirmatory factor analysis (CFA) is a more complex

technique which requires a preconceived structure for the data again based on previous

theory and the model (understanding) developed by the researcher. Concerning the right

choice for this study it was judged that an approach based on an exploratory factor

analysis is more appropriate. Although the Mehrabian and Russell model (PAD and

behavioural variables) is well established in previous research, the model has been applied

in two different settings and the suitability of the model for these settings should be

examined. In addition, especially for the confusion scale an approach based on an EFA is

a proper choice at this early stage of testing. It is essential to note that Schweizer et al.,

(2006) have challenged the overload- similarity and ambiguity tri-dimensional

conceptualisation of confusion by arguing and empirically proving that, at least in the

context of their research, overload and similarity confusion load together/ form one

dimension that was named as the ‘Variety’ factor. This finding further establishes the

necessity to explore the fit of the scale in specific contexts. For all these reasons an


201

approach based on EFA was preferred. Other assumptions and choices around factor

analysis will be explained in more detail in the analysis chapter.

Beyond EFA, correlation coefficients will be examined and other statistical approaches

more widely known as being part of the overall approach of the General Linear Model

(GLM) will be used. ANOVA, factorial ANOVA and regression have been utilised for the

hypothesis testing of this study. ANOVA is called this way because it compares the

variance (variability in scores) between different groups (believed to be due to the

independent variable) with the variability within the groups (believed to be due to

chance). An F ratio is calculated to represent the variance between the groups, divided by

the variance within the groups (Pallant, 2010, p. 214). The use of ANOVA is to compare

the differences between different groups. Multiple regression is again a family of

techniques used to predict relationships between one continuous dependent variable and

one continuous independent variables used as predictor (usually continuous). It is based

on correlation but it allows for a more sophisticated exploration of the relationships of

variables which can be used for prediction and determination (Pallant, 2010, p. 140). The

assumptions that the data need to meet along with the way these techniques serve the

examination of hypotheses will be provided in chapter 8.

7.6. Research Integrity and Quality Criteria

Several criteria to judge the quality of social research have been introduced. The most

prominent approach is to search and establish the three following criteria: replicability,

validity and reliability (Bryman & Bell, 2007, p. 40-43). Generalisability has also been

proposed as central for the conduct of quantitative research (Bryman & Bell, 2007, p.

169). Especially the concept of validity (and the sub-concepts of internal, external,

ecological and construct validity) originated from Campbell and Stanley (1966 as in


202

Calder et al., 1982) and has had a substantial impact on the way researchers think of their

work (Calder et al., 1982). These criteria have been primarily connected with the conduct

of quantitative research and the way these criteria are met by this research will be

described below.

7.6.1. Quality Criteria for Judging the Research as a Whole

According to Bryman and Bell (2007) replicability concerns the possibility of accurate

potential research replication. The replication of studies is a widely debated topic in the

social sciences (Hubbard & Armstrong, 1994; Madden et al., 1995; Easley et al., 2000).

The main issue with replication studies over the years is the perceived lack of importance

and creativity that these have been connected with, resulting in the minimum publication

of replication studies. Notwithstanding, the importance of replication studies should not

be downgraded as such studies add to the establishment of external validity and

additionally, a successful replication promotes confidence as to research and scientific

integrity (Madden et al., 1995; Easley et al., 2000). The way for a study to be replicable is

to provide a detailed explanation of the approach, the research methods, scales, sampling

and measurement and analysis techniques. This methodology chapter along with the four

thorough chapters on the literature review and the subsequent analysis have spelled out all

the necessary details that make this study replicable. The survey instrument

(questionnaire) used is also attached in Appendix 7.

‘Validity is concerned with the integrity of the conclusions that are generated from a piece

of research’ (Bryman & Bell, 2007, p. 41) and has been distinguished into internal,

external, ecological and construct validity. Reliability in turn has been defined as the

extent to which measures are free from errors and capable of yielding consistent results. In

that manner, internal, external and ecological validity concern the research as a whole


203

while reliability and measurement validity mainly concern the structure and items of the

questionnaire.

Internal validity refers to ‘the approximate validity with which we can infer that a

relationship is causal’ (Cook & Campbell, 1979, p. 37). This kind of validity is usually

established through experimental designs where one variable is manipulated and the effect

on another is examined. Cross-sectional survey designs are usually weak in internal

validity and it is also essential to state that this study is not looking to establish causal but

rather functional relationships between the constructs as dictated by the theoretical and

philosophical framework. In this manner this study does not own the faculty of assessing

the criterion of internal validity, which is considered as inappropriate for the specific

endeavour.

External validity ‘examines whether observed relationships should be generalised to and

across different measures, persons, settings and times’ (Calder et al., 1982). External

validity is one of the main reasons that quantitative researchers are striving to generate

representative samples. Calder et al., (1982) based on arguments by Cook and Campbell

(1979) describe however that when a researcher’s interest is mainly theoretical, external

validity can be of little concern, because by trying to achieve external validity other

validity kinds like internal or construct can be compromised (Campbell & Stanley, 1966

an in Calder et al., 1982). It is then a choice of each study whether one of the aims will be

to achieve external validity. This study’s intentions are primarily theoretical but it has

aimed for real-world data (descriptions of every-day situations) from a non-student

sample in order to achieve the most accurate depiction of different groups of the

population and to provide a safe picture of the way people and situations interact. In

addition, in order to achieve the ‘delimitation’ of consumer behaviour, everyday situations

that people confront are used in this study; although these are not placed in the actual


204

settings of occurrence but are rather examined through situational descriptions, these are

expressed in a naturalistic manner which allows consumers to respond to situations in

their own idiosyncratic ways (this goes beyond an experimental design where levels of

variables would be manipulated). Thus although the main aim of this study is clearly

theory testing and extension this study has conformed to the calling of Wells (2001, p.

496) for studies that ‘confront rather than evade’ external validity and achieved some

levels of external validity and generalisability when considering both methodology and

sampling techniques. However, a replication of this study and the examination of

additional market situations (that could go beyond the two examined in this study) would

be necessary in order to argue for the situational generalisability of the findings.

Ecological validity is the extent to which the instrument is capturing the daily life

conditions, opinions, values, attitudes and knowledge base of the study’s population, as

expressed in their natural habitat (Bryman & Bell, 2007, p. 42). The criterion of ecological

validity seems to be satisfied in two ways by the present quest. First of all, data have been

gathered based on real world- retail situations that participants are familiar with.

Secondly, the Mehrabian and Russell (1974) scales have been adopted widely in

consumer settings and have been considered as adequate to capturing consumers’

emotions and behaviours in such settings. The other main scale, the confusion one has

been validated through previous qualitative and quantitative research and has been further

refined through the exploratory/ questionnaire development phase of this research. This

scale has gained acceptance by participants during these initial phases.

Overall, several steps have been adopted in order to assure the most reliable and valid

results, especially considering the online nature of this inquiry (Göritz, 2007). The

questionnaire was meticulously pre-tested for usability; it was as much possible tailored to

the requirements of the sample (this was perceived as a great opportunity to test the


205

responsiveness of especially the PAD scale to online research). It has further been based

on scales tested and established in previous research and extends their applicability in

other situations, it has provided details and information on the two markets and detailed

guidance on the way to complete it, provided information on the approximate completion

time, assured confidentiality and complete anonymity of the data (which is anyway

assured through the quantitative-online nature- information is presented in aggregate form

and no personal information was provided to the research), the main survey data were

collected in three waves including an additional soft-launch phase following which

corrective actions took effect. This research further provided the contact information of

the researcher in case any issues occurred. Finally, a small debriefing was provided at the

end of the questionnaire. Main objectives of this project were explained to the

participants, in an attempt to inform them on the ways their answers would be used.

7.6.2. Specific Quality Criteria for Judging the Quantitative Constructs

Further to those general criteria to judge research, a description of construct validity and

reliability will follow in this section. These aspects are mainly concerned with the

measurement instrument and refer very much to the constructs and items used (full

assessment of the relevant concepts will however be more fully implemented and

measured in the analysis chapter as part of the examination of the psychometric properties

of the scales).

a. Construct Validity

‘Construct validity considers whether or not the operational variables used to observe

covariation can be interpreted in terms of theoretical constructs’ (Calder et al., 1982).

Elements of construct validity are unidimensionality, convergent and discriminant

validity. With respect to unidimensionality, which examines the existence of one construct

underlying a set of items (Steenkamp & van Trijp, 1991), this study has examined whether


206

items of the questionnaire have loaded to their corresponding theoretical construct. In case

this is not so, a further theoretical assessment is necessary in order to examine the

possibility of a new construct that makes theoretical sense (Hair et al., 1998). This

principle has taken effect in this study (see chapter 8) and the theoretical importance of

this finding will be further discussed in chapter 10. Convergent and discriminant validity

require an examination of the strength of the relationships between the variables.

Variables which measure the same construct are expected to have higher correlations

(convergent validity), while those that measure different constructs are having

discriminant validity if their inter-correlations are not too high (De Vaus, 2002).

Exploratory Factor Analysis and an examination of the correlations among the constructs

will be the preferred methods for examining these aspects of validity.

b. Reliability

Reliability is a measurement of the consistency of a measure (Bryman & Bell, 2007).

Generally, the concept is used to examine whether measures produce consistent results

under different circumstances (Peter, 1979, p. 6). In this manner different general classes

of reliability have been proposed (Parasuraman et al., 2004, p. 295-296).

Inter-rater reliability: assesses the consistency of results when tests are taken by different

respondents but using the same methods.

Test-retest reliability: assesses the degree to which scores are consistent from one test to

the next. The same respondents answer the same tests and in the same conditions.

Inter-method reliability: assesses the degree to which scores are consistent when there are

variations in the methods or instruments used.

Internal consistency reliability: assesses the consistency of results across items within a

specific project.


207

Internal consistency will be the preferred and suitable method for this project and will be

examined in the analysis chapter through Cronbach alpha and ITC. Inter-method

reliability will also be discussed as measures will be juxtaposed to previous research

findings in the further analysis chapter (chapter 9).

7.7. Data Preparation for Analysis

For all likert scales a 7-point items scale was preferred as 5-point was evaluated as being

too coarse and thus not being able to capture small differences of the variables (7-point

scales were also preferred in order to increase response variability and minimise ceiling

effects in connection to the PAD variables which are measured on a 9 point scale, e.g.

Zimet, 1988). For the PAD the original nine point bipolar scale was used and a seven

point for Approach-Avoidance ranging from 1 to 7 was preferred. Aminusa which is the

net difference between approach and avoidance (approach minus avoidance) had to be

created as a new variable following the examination of the scales. The statistical software

used for this analysis was IBM SPSS Statistics 20. The main literature used for the

implementation of the analysis was: Hair et al., 1998; De Vaus, 2002; Tabachnick &

Fidell, 2007; Pallant, 2010; Brace et al., 2012. Several other sources were used when more

detailed explanation for a statistical phenomenon was sought.

Nine of the eighteen PAD items were reversed coded before analysis as following:

Table 7.8 Reversed items of the PAD scale

Measured Items Reversed items

Happy 1.............9 Unhappy Unhappy 1.............9 Happy

Autonomous 1.............9 Guided Guided 1.............9 Autonomous

Relaxed 1.............9 Bored Bored 1.............9 Relaxed

Satisfied 1.............9 Unsatisfied Unsatisfied 1.............9 Satisfied

Frenzied 1.............9 Sluggish Sluggish 1.............9 Frenzied

Aroused 1.............9 Unaroused Unaroused 1.............9 Aroused

Controlling 1.............9 Controlled Controlled 1.............9 Controlling

Stimulated 1.............9 Relaxed Relaxed 1.............9 Stimulated

In-control 1.............9 Cared for Cared for 1.............9 In control

Source: this study (following the instructions of Mehrabian & Russell, 1974)


208

The reversing of items resulted to the three items of the PAD pleasure, arousal and

dominance scales scoring from 1= maximal displeasure/ minimal stimulation/ least

dominance to 9= maximal pleasure/maximal stimulation/ maximal dominance. Approach

responses were scored from 1= minimal tendency to approach, to 7= maximal approach,

yielding a range from 1 to 21. Similarly, avoidance responses were scored from 1=

minimal avoidance to 7= maximal avoidance, again yielding a range from 1 to 21.

Approach–avoidance (aminusa) scores were measured as the mean difference between

approach and avoidance; the feasible range was -20 to +20. In addition, the range for

confusion was from 1= minimal confusion to 7= maximal confusion and for market

experience 1= minimal experience to 7= maximal experience.

The items of the high technology situation were placed beneath their corresponding

grocery market items, creating a database of 520 points. In order to facilitate analysis an

additional variable named ‘Market’ (2 values 1=Grocery Market, n=260 and 2= High

Technology Market, n=260) was created.

7.8. Conclusion

This chapter has provided information on the research methodology utilised in order to

answer the research objectives and hypotheses of this inquiry. This research has embraced

a pluralistic conception of marketing research philosophy and has been based on a

philosophical position advocating the principles of intentional behaviourism, which

clearly corroborates the main conceptual edifice behind this thesis; it has utilised a

deductive research design and has defined the purpose of this research as

exploratory/descriptive. It has followed a quantitative design with the emphasis being

placed on a cross-sectional quantitative online survey. The use of the online research

survey has facilitated the examination of research objectives and hypotheses and the


209

exploration of the proposed theoretical frameworks. Multiple pilot tests and assessments

have examined the fit of the questionnaire and the multi-item scales for online research as

data were gathered through the use of an online research panel. The next chapter will

illustrate the process of data analysis used to reach the conclusions of this research.

CHAPTER 8- DATA ANALYSIS

210

8. DATA ANALYSIS

8.1. Introduction

Chapter 8 of this thesis deals with the analysis of the quantitative data collected through

the online questionnaire. The purpose of this chapter is to examine the procedures that

were followed in order to test the research hypotheses proposed in the conceptual

framework. The process used to conduct this data analysis is summarised below in figure

8.1:

Figure 8.1 Process of data analysis

Source: this study

This chapter will commence with an overview of the survey participants’ demographic

characteristics, the way data were examined and managed, the results of descriptive,

univariate analysis and will then proceed to the results of various bivariate and

multivariate statistical tests.

1. Explore Data and Preliminary Analysis

Sample characteristics

Missing data

Outliers (univariate and multivariate)

Normality (univariate and

mulitvariate)

Descriptive statistics

2. Establish measurement scales

EFA

Reliability

Scales’ descriptive statistics

3. Hypotheses testing BPM-E/ BPM-I

Test whether variables will

differentiate as predicted by the model

(ANOVA)

Test overall model

Relationships (correlation)

Degree of the relationship (multiple

regression)

Interaction effect (factorial ANOVA)

The effect of market (experience) and

confusion (factorial ANOVA)


211

8.2. Sample Demographic Profile

The respondents of this research were members of an online research panel. In total 264

fully completed (with the occasional missing data) questionnaires were provided by the

research panel agency. The requirement was for a diverse sample which would be

representative of the characteristics of the UK population. During the soft launch 2,000

invites were sent in order to collect 100 fully completed questionnaires. The second main

phase included 6,000 invites and during the last complementary phase 1,400 invites were

sent. In total 321 questionnaires were collected (this excludes participants who were

excluded in early stages due to this quota sampling) of which 264 were judged as been

usable, passing the research panel quality control (less than 10% missing data and

completion time of more than 6 minutes). All 321 questionnaires usable or not based on

the research agency judgement were provided to the researcher for further elaboration.

One week time period was offered in order to allow sufficient time for initial analysis in

case any further quality problems were identified (many missing data, half completed, not

good sampling etc.). All questionnaires were checked for set response bias and four were

discarded since these respondents had completed the same or very similar answers in the

scale items. This resulted in a usable number of 260 respondents/questionnaires.

In total eight questions of this questionnaire concerned questions on participants’

demographic profile. These same demographic questions were used by the research

agency for quota purposes. More specifically information regarding age, gender, higher

completed educational level, ethnic group, working status, household size and finally

grocery and PC/Laptop shopping habits were collected.

123 participants identified themselves as males (47.3%) and 137 as females (52.7%).

They further identified their ethnic group as following: 89.2% white, 2.3 mixed, 6.5%

Asian (or Asian British), 1.2% as black and 0.8% as other (e.g. Arabic). A majority of


212

respondents (60%) lives in larger households (more than 2 persons). However, following

the trend towards smaller households 40% live in either 1 person (15%) or 2 person

households (25%). Useful information concerns the channels that respondents typically

use in order to shop for their groceries and technological products. More specifically,

these statistics justify the choice to include the online element to the description of the

PC/Laptop situation as 49.2% (128 respondents) of the respondents use this channel to

shop for technological products, while the same statistic is much lower in the case of

grocery shopping– 15.4% (40 respondents).

A complete description of the sample’s socio-demographic profile is provided in table 8.1


213

Table 8.1 Sample demographic profile

n=260

n %

Gender Males 123 47.3%

Females 137 52.7%

Age 18–24 41 15.8%

25–34 70 26.9%

35–44 72 27.7%

45–54 48 18.5%

55–64 23 8.8%

65+ 6 2.3%

Higher completed

education

GCSE/ A-levels 101 38.8%

Vocational or Technical School 30 11.5%

Higher Education 91 35%

Postgraduate Degree 35 13.5%

Other 3 1.2%

Ethnic Group White 232 89.2%

Mixed 6 2.3%

Asian (Asian British) 17 6.5%

Black (Black British) 3 1.2%

Other 2 0.8%

Working Status Employed full-time (30+ hours per week) 109 41.9%

Employed part time (less than 30hours per

week)

37 14.2%

Unemployed 24 9.2%

Student 26 10%

Retired 17 6.5%

Self-employed 11 4.2%

Housewife/husband 36 13.8%

Household Size 1 person 39 15%

2 person 65 25%

More than 2 persons 156 60%

Channel for buying

groceries

In store 40 84.6%

Online 220 15.4%

Channel for buying

PC/laptops

In store 132 50.8%

Online 128 49.2%

Source: this study

Regarding the answer to the question, how frequently do you shop for either technology

or grocery products the average answer for groceries was 2-3 times a month (113 out of

the 260 respondents) and the average answer for PC/Laptops was between once a year

and once every two years (154 out of 260 respondents). This has been perceived as an


214

additional proof of the lesser experience characterising shopping for high technology

products.

8.3. Preliminary Analysis

8.3.1. Missing Data

Missing data is a pervasive problem in any data analysis (Tabachnick & Fidell, 2007, p.

62; Baraldi & Enders, 2010). Missing data can be the result of data entry (in cases that the

questionnaires take the ‘pen and paper’ form and data entry is manual) and software

malfunctions (when online software is used). It can further be attributed to participants

becoming bored, recalcitrant or simply confused from the structure of the questionnaire or

the mere amount of questions and thus fail to answer some items. Missing data can pose a

problem, both regarding their proportion but also depending on the ‘mechanism’ of

missing data (Baraldi & Enders, 2010). The ‘mechanism’ of missing data refers to the

classification system more frequently used today to describe the pattern behind missing

values (Little & Rubin, 2002 as in Baraldi & Enders, 2010). These patterns have been

described as following: Missing Completely At Random (MCAR), Missing At Random

(MAR) and Missing Not At Random (MNAR). Actually according to Tabachnick and

Fidell (2007, p. 62) the pattern of missing data is more important than the amount missing.

Missing data which are scattered randomly pose less serious problems because there are

statistical valid ways to diminish their effect (Baraldi & Enders, 2010; Hair et al., 1998)

while non-randomly distributed missing data (even if less in number) can be an issue for

the genarilisability of results and require further elaboration to examine the reasons of

occurrence, occasionally even require the deletion of variables that seem to cause or

present the non-random distribution of the missing values (De Vaus, 2002, p. 176).


215

a. Handling of Missing Data

There are several methods that have been proposed to handle the problem of missing data.

There are two big categories, one category involves the deletion of cases, namely list-wise

or pair-wise deletion and the other is the estimation of valid values for the replacement of

missing data, namely mean imputation and regression-based substitution (De Vaus, 2002,

p. 176-177; Allison, 2009). Listwise deletion involves the exclusion of cases or

respondents with missing items from the analysis. That means that if any case has missing

data, this case is omitted from the computation. The problem with this approach is that it

can easily result to the loss of lots of data and information, especially when the number of

missing data is large. Pairwise deletion involves the use of all cases that have no missing

items even if those cases have missing values on other variables being used in the

analysis. When using this approach we end up with a situation where each coefficient may

be based on a different number of cases (De Vaus, 2002, p. 176; Allison, 2009, p. 76)

were results produced might be unappealing.

The second group of missing data ‘treatment’ is the replacement of missing values with

some estimated scores. These scores are the result of respondents’ pattern of answers in

other questions and are used to replace the missing data. The most common of these

approaches is mean imputation and regression based substitution. Mean imputation

technique advocates the replacement of the missing value with the best guess for that

missing value which is the measure of the central tendency of that variable or the mean for

interval level variables (De Vaus, 2002). This approach has been however, widely

questioned because it can easily decrease the variability in the dataset and can result in

reduction of correlations between the variables (Malhotra & Birks, 2007). The final

approach of regression based substitution is a more complex approach using regressions

based on cases with complete data to predict values for the incomplete cases (Allison,


216

2009). The positive thing about it is that the imputed value is at least somehow conditional

to the rest of the information about the case. It has generally been described as producing

accurate estimates and has been seen as the recommended approach when considering the

aforementioned traditional approaches to missing data handling (Allison, 2009, p. 87).

On the grounds of the sections on missing data and handling of missing data, any

decisions on the most suitable missing data handling method for the present research, took

into consideration both the amount and ‘mechanism’ of missing data and the

aforementioned strengths and weaknesses of the methods. Specifically, in terms of

numbers this dataset has not suffered significantly from missing data, especially when

considering the main variables like emotions and confusion. Overall missing items were

dispersed across cases. One alarming issue is that questions on socio-demographics

present a bigger however not extreme indication of missing data. Specifically, overall only

34 cases are counted as missing data in this database. Appendix 5 shows the frequencies

and percentages of missing responses for each item (the items are presented collectively

for both markets).

In terms of the missing data mechanism, some handbooks of statistics usually advocate for

the creation of two groups- missing and non-missing items groups- and the performance

of t-tests accordingly (De Vaus, 2002) in order to spot any differences among groups.

Recently the Roderick J. A. Little's chi-square statistic (Little, 1988) for testing whether

values are missing completely at random (MCAR) is printed as a footnote to the EM

(Expectation Maximisation) matrices of the MVA (Missing Value Analysis) of SPSS. For

this test, the null hypothesis is that the data are missing completely at random, and the p

value is significant at the 0.05 level. If the value is less than 0.05, then the data are not

missing completely at random and this might be a problem for data analysis. If the test

indicates that data are not missing at random then further analysis would be necessary in


217

order to assess the extent of the problem. Little’s chi square statistic is a chi-square

statistic and for this study it is reported as:

Overall data (including socio-demographics): χ² (295, N = 520) = 277.161, p = .765

(Grocery Market: χ² (222, N = 260) = 175.564, p = .991 and

PC/Laptop Market: χ² (221, N = 260) = 233.080, p = .276).

On the grounds that the dataset suffers from a very small number of missing data, the

Little’s MCAR (Missing Completely At Random) test has indicated that data are indeed

missing completely at random and considering that the regression based substitution is

generally the proposed method (Allison, 2009, p. 87), this study has utilised this approach

for handling the missing cases11

.

8.3.2. Outliers (Univariate and Multivariate)

The second issue of importance in the preliminary analysis is the existence of outliers in

the database. An outlier is a case with an extreme value on one variable widely separated

from the rest of the data (this is called a univariate outlier) (Howell, 2007, p. 20) or with a

strange combination of scores on one or two variables (named a multivariate outlier)

(Tabachnick & Fidell, 2007, p. 72). In that sense, outliers are observations that are

inconsistent with the remainder of that data. The problem with outliers is that due to their

extreme values these can have an increased impact and distort statistical values. They can

once again be the result of error in data entry or a measurement error. However, one

should not neglect the case that an outlier is a legitimate and correct value that just

happens to be extreme, however it represents reality. This is the reason that deciding

11 This research acknowledges the superiority of the modern techniques of missing data handling, like

maximum likelihood technique and multiple imputation, as explained by their advocates (Allison, 2009;

Baraldi & Enders, 2010). The complexity of these methods, the requirement for specialised statistical

software (SPSS does not support these techniques) along with the minor impact (very little cases and no sign

of missing data bias) that missing data seems to have in this dataset have resulted in a balanced and

informed choice to follow one of the traditional techniques.


218

between deleting and retaining these cases is a difficult endeavour (Hair et al., 1998, p.

67; Tabachnick & Fidell, 2007, p. 72).

Following instructions by Pallant (2010, p. 58-62) and Brace et al., (2012) univariate

outliers have been examined using box plots and histograms for all individual items of the

questionnaire. In accordance with Pallant (2010) all trimmed means for all items were

checked. Trimmed means represent the value of the mean when the top and bottom 5% of

the cases are removed. This is a good indication of the effect of the outliers. Trimmed

means are especially important in this study where Anova (mean comparisons) will be

frequently used. The investigation of the univariate outliers did not revealed any

significant differences between the means and the trimmed means and any outliers were

within the expected range for the variables. An overall observation that could guide future

research is that most univariate outliers were identified with the arousal items like for

example Frenzied-Sluggish and Dull-Jittery. This finding might be an indication of what

Donovan et al. (1994) also pointed in their research. They based their argument on

anecdotal feedback from their participants who indicated that those two items plus the

aroused-unaroused item could not really correspond to the feelings experienced in a retail

setting.

The criterion for multivariate outliers is Mahalanobis distance at p<.001, assessed as a chi

square (χ²) with degrees of freedom equal to the number of variables (Tabachnick &

Fidell, 2007, p. 99). The calculation of the Mahalanobis distance was performed using the

regression process. Specifically, the ID number of the cases was selected as the dependent

variable and the 38 items of the questionnaire as the independent variables. For 38 items

the criterion of χ² at p<.001 is 73.402 (table of critical values of χ² as in Tabachnick &

Fidell, 2007, p. 949). Any value with Mahalanobis distance above this threshold is

considered a multivariate outlier. The analysis revealed seven cases (out of 520) of


219

multivariate outliers. The values for five cases are reported by SPSS: Case 164 with

Mahalanobis distance 127.79, case 424 with a value of 127.48, case 296 with a value of

122.34, case 159 with a Mahalanobis value 113.14 and case 316 with a value of 100.32.

All cases of univariate and multivariate outliers were checked elaborately and outliers

could not be attributed to a data entry or a re-coding problem. In accordance with several

researchers that argue (Hair et al., 1998, p. 66) that in social sciences outliers are usually a

valid representation of reality (when there are no obvious data entry errors), outliers were

retained in this dataset for further analysis12

.

8.3.3. Normality (Univariate and Multivariate)

Normality refers to the distribution of variables. Specifically, it is a way to describe the

way that a distribution corresponds or fits to the normal distribution (Howell, 2007, p. 67).

It is accessed through the measurement of skewness and kurtosis (Tabachnick & Fidell,

2007, p. 79) of each variable or the composite scores. Skewness is the degree of

asymmetry of a distribution (Howell, 2007, p. 27), referring to the position of the mean.

Kurtosis refers to the shape of the curve mainly in terms of concentration of the values

around the centre and the two tails (Howell, 2007, p. 29). The normal distribution (widely

named as Gaussian) has a mean, a skewness and a kurtosis of 0 (zero) and a standard

deviation of 1 (Howell, 2007, p. 73).

As in the case of outliers normality can also be described in terms of univariate and

multivariate normality.

12 In order to reach this decision analysis was re-run without the identified multivariate outliers. The results

did not vary significanlty to the ones produced in case these outliers remain in the dataset. Outliers were

then retained in order to provide a picture as close to the data collected as possible.


220

Univariate normality implicates the examination of the distributions. Univariate normality

is an assumption underlying the performance of several statistical techniques and as such

it is a focal point for the researcher in order to make the choice of certain analysis

techniques rather than others. For this reason it is a matter of great importance and it is

included in this section of data management.

Specifically, the category of parametric statistical tests (e.g. pearson’s r correlation

coefficient) have been designed on the assumption that data correspond to the normal

distribution while non-parametric tests (e.g. spearman’s rank correlation coefficient), are

not based on this assumption and should be the choice when data are not normally

distributed (Micceri, 1989). The use of one approach over the other (parametric versus

non parametric) is usually not fatal in terms of identifying relationships between the data.

It usually however means that when the inappropriate technique is used some estimates

will be under- or over- determined depending on the approach (Tabachnick & Fidell,

2007).

Normally distributed data is a pragmatically uncommon occurrence in social sciences

(Micceri, 1989; Pallant, 2010) but non-parametric tests are not to be considered

appropriate immediately and without further judgement. According to Micceri (1989, p.

161) :‘adequate research is available to suggest that most parametric statistics should be

fairly robust to both alpha and beta given light tail weights and moderate

contaminations’.

Many parameters should then be taken into consideration when choosing the appropriate

techniques in connection with normality. Three of these parameters are considered in this

study: 1) How far the data differ from the normal distribution? Analysis should be

performed by non-parametric tests only if the data are abnormally distributed, that is


221

according to Brace et al., (2012, p. 85) severely skewed. Severe kurtosis can also result in

an underestimation of variance but it is not fatal for larger samples. 2) Sample size is then

another crucial element that the decision should be based with most statistical books

(Tabachnick & Fidell, 2007 as in Pallant, 2010) arguing that a sample of 200 cases or

more is sufficient in order to overcome the issues faced by non-normal distributions. 3)

Finally one additional parameter is the availability of non-parametric alternatives. Several

statistical techniques, like Factorial Anova (GLM) for example, have no non-parametric

alternative (Pallant, 2010, p. 241) and in this case it lies with the researcher to decide

whether the use of this technique is justifiable or it would be advisable to find an

alternative statistical method. The sample size along with the consideration of the fine

levels of skewness and kurtosis, nonchalantly allows the use of parametric tests in this

study.

The second kind of normality, multivariate normality depends on univariate normality (De

Vaus, 2002, p. 344) and is an indication of whether each variable and all linear

combinations of variables are normally distributed. It is mainly an issue in such

techniques as multiple linear regression and it is examined in cases such statistical

approaches are to be used by consulting the distribution and independence of the residuals

(Tabachnick & Fidell, 2007, p. 78) -see the residual plots of the multiple regressions of

this analysis where multivariate normality is assessed along with the assumption of

heteroskadasticity.

a. Assessing Univariate Normality

This research has used several ways to access normality. The Kolmogorov-Smirnov test

of normality (SAS Institute, 1985 as in Micceri, 1989) can be used to access a

distribution’s normality, however as it adopts very stringent assumptions it usually finds

distributions not to be normal at the .01 alpha level. For larger samples sizes the


222

likelihood of rejecting a variable that deviates only slightly from normality increases

(Howell, 2007). Examination of the actual parameters of skewness and kurtosis

(descriptive statistics) and graphical inspection of the data in terms of histograms, stem-

and-leaf plots and Q-Q plots assisted the examinations of the distributions in this study.

Among the ways to deal with not normally distributed date are the use of non-parametric

tests as described before, the use of trimmed means (for heavy-tailed distributions) and

also the transformation of the values which is applied in order to transform their

distribution (Howell, 2007, p. 317-323). Data transformations usually involve such

processes as transforming the original data into their logarithms or square roots (Howell,

2007, p. 317-323). It is not unusual to see such transformations although considering that

after performing the transformations, researcher is not working any more with the real

data, caution should be taken in the results and explanations extracted. Generally, it is

suggested that such extreme measures should not be taken in social science data if

distributions are reasonably distributed with only a few outliers (Howell, 2007, p. 323).

8.4. Descriptive Analysis

This section will provide the descriptive analysis of the individual items used in this study

(descriptive statistics for the composite scales will be provided later in this analysis).

Descriptive statistics provide a summary of what a research has found. Means, standard

deviation and measures of normality (skewness and kurtosis) of all the items will be

provided. All negatively worded items have been reversed prior to these calculations.

8.4.1. Means (Sd)/ Skewness/ Kurtosis of Individual Items

The following table 8.2 summarises the descriptive statistics of the individual items of this

study. Specifically, mean and standard deviation, skewness and kurtosis values are

presented for all items.


223

Table 8.2 Descriptive statistics of individual items

Item Mean (SD) Skewness Kurtosis

P1. Unhappy- Happy 6.55 (1.8) -.534 -.178

P2. Annoyed- Pleased 6.21 (1.9) -.338 -.471

D1. Guided-Autonomous 4.99 (1.6) .079 -.001

P3. Bored- Relaxed 6.28 (1.9) -.412 -.461

A1. Excited- Calm 4.78 (1.9) .242 -.542

P4. Unsatisfied- Satisfied 6.34 (1.7) -.251 -.336

P5. Melancholic- Contented 6.05 (1.8) -.244 -.098

P6. Despairing- Hopeful 6.36 (1.7) -.412 -.080

A2. Frenzied- Sluggish 5.18 (1.4) .122 1.49

D2. Awed- Important 5.37 (1.3) .343 .817

A3. Dull-Jittery 5.07 (1.3) -.051 1.74

A4. Unaroused- Aroused 4.63 (1.9) -.332 -.143

D3. Controlled- Controlling 5.10 (1.7) .199 -.214

A5. Relaxed- Stimulated 5.51 (1.8) -.213 -.294

D4. Influenced- Influential 4.70 (1.5) .097 -.120

D5. Cared-for- In control 5.89 (1.5) -.323 .109

A6. Sleepy- Wide-awake 6.19 (1.8) -.390 -.242

D6. Submissive- Dominant 5.46 (1.4) -.059 .438

AP1. Time spent 3.96 (1.2) .350 .362

AV1. Try to leave 2.70 (1.7) .658 -.575

AP2.Enjoy exploring 4.44 (1.6) -.311 -.627

AV2. Avoid others 3.33 (1.8) .226 -1.009

AP3. Feel friendly 3.60 (1.6) .179 -.542

AV3. Avoid looking around 2.59 (1.7) .755 -.404

S1. Difficult to spot new products 4.32 (1.6) -.403 -.553

S2. Difficult to detect differences 4.39 (1.6) -.432 -.443

S3. Hard to distinguish 4.09 (1.5) -.107 -.662

S4. Made by the same manufacturer 4.13 (1.5) -.122 -.526

O1. The harder it gets to choose the best 4.17 (1.7) -.127 -.872

O2. Many brands-feel confused 4.14 (1.6) -.194 -.710

O3. Many shops to shop at 3.86 (1.7) .010 -.899

O4. Too much information 3.67 (1.7) .152 -.749

O5. Too many products to choose from 4.02 (1.7) -.020 -.616

AM1. The comparison of brands is difficult 4.03 (1.6) -.145 -.647

AM2. I feel uninformed 3.49 (1.4) .211 -.222

AM3. I get vague information 4.11 (1.5) .040 -.403

AM4. Uncertainty about product

characteristics

3.90 (1.6) -.022 -.580

ME1. Market experience 5.08 (1.4) -.686 .278

Source: this study


224

Distributions are in their majority platykurtic (negative kurtosis), indicative of relatively

flat distributions (Pallant, 2010). Largely, the items13

do not indicate extreme skewness

and kurtosis values, the sample of this study is far above the 200 cases proposed

(Tabachnick & Fidell, 2007) -260 cases for each situation and 520 when taken

collectively- allowing the confident use of parametric statistics; this analysis has valid

statistical reasons to follow the norms of relevant studies that used parametric statistics for

the investigation of the relevant hypotheses.

8.5. Psychometric Properties of the Scales

This study has used multivariate measurements, more widely known as summated scales

or scales for measurement purposes. The practice of using summated scales is widely

utilised in marketing and psychology and the objective is to avoid the use of only a single

variable to represent a concept. Instead summated scales are comprised of several

variables joined together in a ‘composite measure’ to represent that concept (Hair et al.,

1998, p. 10). As explained in the methodology chapter there are recent voices (Drolet &

Morrison, 2001; van Birgelen et al., 2001; Fuchs & Diamantopoulos, 2009) advocating

that the use of single items in specific research cases is not only justified but in addition

desirable, and such voices should not be disregarded. However, this study has used scales

for its main constructs (apart from the measurement of market experience) where each

single variable of the scale is supposed to characterise differing facets of the concept and

altogether to represent a more coherent perspective of the concept under investigation

(researcher is basing findings on the ‘average’ response to a set of related responses).

Specifically, psychometrics is the field of study examining the theory and technique of

psychological measurement, which is especially concerned with the construction and

13 Items were also examined on a split by market basis and the results indicated approximately similar

results. For reasons of clarity of presentation only the overall items are presented here.


225

validation of measuring instruments (Borsboom, 2005). Methodology chapter has defined

the basic issues when judging the quality of a measurement scale and these are:

dimensionality, measurement validity and reliability. The following two sections will deal

with dimensionality and reliability.

8.5.1. Dimensionality

Dimensionality of the scales is empirically assessed by factor analysis (Hair et al., 1998).

Factor analysis (FA) is the statistical approach used to analyse the interrelationships

among a large number of variables and to explain these variables in terms of their

common underlying dimensions (factors). By providing the necessary estimate of the

structure of the variables considered, this method has acted as a basis for creating and

assessing the quality of summated scales (Hair et al., 1998, p. 14). It simplifies the

correlations and reveals important information by a new and smaller set of variables. Its

purpose is then the identification of the structure of the data.

The first issue to consider when performing factor analysis is the distinction between

Exploratory (EFA) and Confirmatory (CFA) Factor Analysis. This distinction has been

described before (chapter 7- Methodology) and the choice of the exploratory approach

(EFA) in this research has been justified. Another issue to mention is that EFA can be

performed without assuming normality of the data and in that manner the assumption of

normality is not in force here (Tabachnick & Fidell, 2007, p. 613).

This study has then preferred EFA and specifically the method of Principal Component

Analysis and mainly followed guidelines by Brace et al., 2006; Tabachnick & Fidell, 2007

and Pallant, 2010. Especially Tabachnick & Fidell (2007, p. 608) propose that the steps in

PCA include the measurement of the variables, examining the correlation matrix for all

individual items for both extreme correlations (>.90 or >.80 according to De Vaus, 2002,


226

p. 116) or items that have no correlations with any other items, extracting a set of factors

and determining the number of factors and ‘rotating’ the factors in order to increase

interpretability. The final step is the interpretation of the results based on theoretical and

practical considerations. The rotation of factors is a process by which the solution is made

more interpretable. There are two general classes of rotation: orthogonal and oblique. If

rotation is orthogonal (the proposed method here is usually varimax rotation- see all

aforementioned sources) then interpretation is the result of the produced ‘loading matrix’

which indicates which observed variables are correlated with each factor. At the other

end, oblique rotation implies and allows for the factors extracted being correlated (Zimet

et al., 1988) and produces several additional matrices. The structure matrix indicates

correlations between factors and variables and the pattern matrix reveals the unique

relationships between the factors. Following oblique rotation, the meaning of the factors is

ascertained from the pattern matrix (Ho, 2006, p. 220; Tabachnick & Fidell, 2007, p. 609).

The interpretation of the matrices depends on the understanding of the underlying

dimensions that unifies the group of variables that load on this factor. These loadings are

usually indications of the correlations between the variables and the factors these underlie.

Tabachnick & Fidell, (2007, p. 649) identify that a rule of thumb for the items’ loadings

would be that loadings of .32 and above are interpreted. Hair et al., (1998, p. 112, table

3.2) offer a more standard approach which takes into consideration the sample size of the

study. To this latter source any sample size over 350 (n=350) requires a loading of .30 for

significant interpretation. In the case of this research this required factor loading is then

very similar to the rule of thumb proposed by Tabachnick & Fidell (2007). However the

required factor loading increases significantly for smaller sample sizes (a sample size of

100 requires a factor loading of at least .55 in order for the item to be considered a valid


227

component of the scale). Overall, the greater the loading the better because this indicates

that the item is a pure measure of the specific factor.

Further to the aforementioned issue, according to De Vaus (2002, p. 187) caution should

be drawn to one theoretical and another practical fact. Firstly, since the factor analysis

solution is based on the correlations between the variables, regardless of the variables

used a set of underlying factors might be produced- whether they make theoretical sense

or not. The understanding developed then necessitates judgement on theoretical basis.

Secondly, the statistical suitability of data for factor analysis should be verified. In order

to assess whether a set of variables is suitable for factor analysis further to be based on the

conceptual definition, the need to examine the matrix of correlations along with the

Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett‘s test of

Sphericity is crucial (Pallant, 2010; Brace et al., 2012). De Vaus (2002, p. 188) suggests

that KMO values above 0.7 indicate adequate correlations for the factor analysis to be

performed and values between 0.5 and 0.69 indicate that more care should be taken when

analysing the factors produced. Values below 0.5 clearly indicate unsuitable data.

However, Pallant, 2010 suggests that a KMO value of 0.6 and above and a significant

Bartlett‘s test of Sphericity (p<0.05) are enough for the factor analysis to be considered

appropriate.

In order to access the overall suitability of the data and verify that variables like for

instance confusion and dominance do not load together an overall Principal Component

Analysis was performed without rotating the data.

These statistics are presented in the table below:


228

Table 8.3 KMO and BTS measure of sampling adequacy. All items

Kaiser-Meyer-Olkin measure of sampling adequacy .912

Barlett’s test of sphericity Approx. Chi- Square

Df

Sig.

10679.886

703

.000

Source: this study

Kaiser‘s criterion or eigenvalue rule of 1.0 or more was used in all cases to assist in the

decision concerning the number of factors to maintain. The findings are shown in Table

8.4.

Table 8.4 Total Variance Explained

Component Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings

Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative %

1 9.401 24.741 24.741 9.401 24.741 24.741 7.081 18.634 18.634

2 6.053 15.929 40.669 6.053 15.929 40.669 5.996 15.778 34.412

3 2.302 6.057 46.727 2.302 6.057 46.727 2.665 7.013 41.426

4 1.779 4.682 51.409 1.779 4.682 51.409 2.511 6.608 48.033

5 1.396 3.675 55.084 1.396 3.675 55.084 1.906 5.015 53.049

6 1.214 3.196 58.279 1.214 3.196 58.279 1.586 4.174 57.223

7 1.171 3.083 61.362 1.171 3.083 61.362 1.455 3.828 61.051

8 1.100 2.894 64.256 1.100 2.894 64.256 1.218 3.205 64.256

Source: this study

Extraction method: Principal Component Analysis

Overall, the variance explained by all of the items is 64% and the items loaded correctly

to their responding factors. An alarming finding however was that although confusion did

not load together with any of the other constructs, all items of different kinds of confusion

loaded together on the first factor of this factor analysis, apart from item AM2. In order to

deal with this issue and examine the structure of emotional- behavioural and confusion

variables more accurately separate factor analysis (with varimax rotation) was run for the

emotional items as these items are expected to be orthogonal and a further factor analysis

with oblique rotation was run for confusion. Oblique rotation is frequently used when

factors are expected to be correlated and has been used in the past when multidimensional

scales are examined (Zimet, 1988). In that case the factors underlying confusion could

potentially indicate high correlations as these underlie a common theme, that of confusion.


229

Principal Component Analysis for emotional- behavioural variables (Varimax

Rotation)

Table 8.5 KMO and BTS measure of sampling adequacy. Emotional variables



Df

Sig.

4099.638

153

.000

Source: this study

Table 8.6 Factor Analysis (varimax rotation) of the affective and behavioural

measures

Affective Behavioural

Item Factor 1:

Pleasure

Factor 2:

Dominance

Factor 3:

Arousal

Factor 1:

Approach-Avoidance

P2. Annoyed-

Pleased .846

P1. Unhappy-Happy .828

P4. Unsatisfied-

Satisfied .822

P3. Bored-Relaxed .815

P6. Despairing-

Hopeful .799

P5. Melancholic-

Contented .748

D3. Controlled-

Controlling .747

D4. Influenced-

Influential .724

D1. Guided-

Autonomous .718

D5. Cared for- In

control .591

D6. Submissive-

Dominant .570

D2. Awed-Important .409 .457

A2. Sluggish-

Frenzied .683

A5. Relaxed-

Stimulated .665

A3. Dull-Jittery .649

A1. Calm-Excited. .593

A4. Unaroused-

Aroused .577

A6. Sleepy-Wide-

awake .537 .546

AV3. Avoid looking

around

.843

AV2. Avoid other .762


230

people

AV1. Try to leave .748

AP2. Enjoy

exploring

-.312 .804

AP1. Time spent .742

AP3. Feel friendly

and talkative

.666

Explained Variance 26.9% 14.8% 14.2% 44.5% 17.5%

Total 56% 62.14%

Source: this study


Note. The items are sorted by size. Loadings less than .32 do not appear in order to facilitate the

presentation.

These results indicate that overall the items load well on their corresponding factors. All

of the items of the first factor ‘pleasure’ indicate the best consistency and loadings greater

than 0.70. This factor accounts for 26.9% of the variance.

Regarding the second factor ‘Dominance’ one of the items ‘D2. Awed-Important’ that

should conceptually load on this factor has loaded almost equally on both the first and

second factor and thus it is one of the items considered for elimination. The same problem

is evident with one of the items of the third factor ‘Arousal’. This item which should

conceptually belong to arousal again loads equally on both the first ‘pleasure’ and the

third factor ‘arousal’. Item ‘A6. Sleepy-wide-awake’ is also considered for elimination.

The problem with items that so clearly and equally load on two factors is that these

artificially inflate the correlation between the constructs after these have been summated.

A second factor analysis was conducted after the decision to eliminate these two items.

The new factor analysis explained 57.48% of the variance and all of the remaining 16

items loaded on their respective factor.


231

Principal Component Analysis for the confusion scale (Oblique Rotation)

Table 8.7 KMO and BTS of sampling adequacy. Confusion



Df

Sig.

4450.541

78

.000

Source: this study

Table 8.8 Factor Analysis (oblique rotation) of the confusion measures

Item Factor 1:

complexity

Factor 2:

similarity

O4. Too much information .817

AM4. Uncertainty about product characteristics .750

O5. Too many products to choose from .716

O2. Many brands-feel confused .710

AM1. The comparison of brands is difficult .695

O3. Many shops to shop at .691

AM2. I feel uninformed .560 .364

AM3. I get vague information .555

O1. The harder it gets to choose the best .530 -.386

S2. Difficult to detect differences -.855

S3. Hard to distinguish -.815

S1. Difficult to spot new products -.794

S4. Made by the same manufacturer -.671

Explained Variance 47.28 16.88

Total 64.16

Source: this study.


Rotation Method: Oblimin with Kaiser Normalization

Following the results of this factor analysis with oblique rotation, it is evident that the

scale on confusion indicates two factors (as the items load on two factors only). The

second factor clearly indicates the factor of similarity (all items conceptualised as

measuring similarity have clearly loaded there). Items of both the overload and ambiguity

confusion loaded together on the first factor. The common characteristic of these two

kinds of confusion is that these create a complex environment (in contradiction to the

homogeneity caused by product similarity) for consumers. Based on this theoretical

argument this factor was kept as it is, as the underlying interpretation has been revealed

and has been perceived as indicating complexity confusion. Two of the item AM2 and O1

were dropped. The problem with these items is that these load on both factors. As the


232

factors are already expected to have high inter-correlation, these were dropped to reduce

this issue. Keeping them would not be an issue for the factor itself as for the higher

correlation with the second factor. Thus 4 items are measuring similarity and 7 items

complexity confusion in this research. The implications of this finding for the study of

confusion will be further discussed in the concluding chapter (chapter 10).

8.5.2. Reliability

Reliability is used in this study in order to measure the internal consistency. It will then be

assessed by determining the overall consistency of the measures. Cronbach’s alpha is the

preferred coefficient used to measure the internal consistency and this coefficient has been

established as one of the first measurements calculated to assess the quality of a

measurement (Churchill, 1979; Steenkamp & van Trijp, 1991; Cortina, 1993). However

reliability assessment has been the result of an interaction with the technique of EFA

(Exploratory Factor Analysis), which was used to determine the constructs.

The cut off value of alpha for accepting a construct is 0.5 (below that the construct can be

described as unacceptable). However a more up to standard a > 0.7 is proposed and 0.6

has been described as acceptable especially for exploratory research (Hair et al., 1998, p.

118). In addition, as part of alpha value examination another indication will be used to

examine and establish reliability in this study, this is the Item-Total-Correlations (ITC).

The final scale for any construct should not include any items with items score lower than

.30 (Leech et al., 2005, p. 66). The higher the item-total-correlation index for an item

(especially moderately high to high correlations around .40 and over) indicates that the

item will be a good component of a scale. All the decisions to drop/ keep items should be

of course negotiated based on the results of the factor analysis, the findings of ITC, alpha

values, and judgement by the researcher based on previous research and theoretical

grounds. The alpha coefficients of this study are presented in table 8.9. Along with the


233

overall evaluation (N=520), a presentation of the alpha coefficient per situation (N=260)

is also provided:

Table 8.9 Reliability (cronbach's alpha) for the scales of this study

Variables Number

of items

Cronbach’s

alpha (overall)

Cronbach’s alpha

(per market)

Pleasure 6 .923 Grocery .919

High- technology .919

Arousal 5 .661 Grocery .690


Dominance 5 .731 Grocery .726


Approach 3 .633 Grocery .653


Avoidance 3 .727 Grocery .722


Aminusa AP-AV N/A Grocery N/A

High- technology N/A

Similarity

Confusion 4 .916

Grocery .915


Complexity

Confusion 7 .903

Grocery .877


Market

Experience 1 N/A

Grocery N/A

High- technology N/A

Source: this study

Because of negative average covariance among items, the behavioural factor cannot yield

a single alpha coefficient (Foxall & Soriano, 2011). Thus alpha coefficient is provided

distinctively for the Approach and Avoidance constructs.

In summary and in accordance with the proposed cut-off point of 0.60 this study has

achieved good internal consistency, with all items alpha coefficient ranging from above

0.60 to 0.90. This is consistent at an overall and situational level.

In addition, the levels of ITC were satisfactory for all items. Only 2 items had ITC lower

than .40. Item AP1 (time spent) with ITC of .388 and item A1 (calm-excited) with ITC

.336. However because both values are above 0.3 and in addition the cronbach’s alpha

value of the specific scales these items are part (meaning approach and arousal) was


234

actually dropping rather than increasing when removing these items (for Approach alpha

coefficient if item deleted would be .614 and for Arousal alpha if item deleted .655) and

based on previous research, these items were not removed from the summated constructs.

8.6. Hypothesis Testing

Following data examination, this section will then move to test the hypotheses of this

research project.

Before proceeding to the specific tests a table (table 8.10) with the descriptive statistics of

the composite measures (means (Sd)/ skewness/ kurtosis) is provided:

Table 8.10 Descriptive statistics of composite variables

Item Mean (SD) Skewness Kurtosis

Pleasure 37.79 (9.18) -.258 -.180

Arousal 25.16 (5.53) -.084 .773

Dominance 26.17 (5.37) .173 .717

AP 11.99 (3.33) -.052 -.154

AV 8.61 (4.16) .566 -.146

AP_AV (Aminusa) 3.38 (6.37) -.438 .350

Similarity confusion 16.93 (5.49) -.289 -.276

Complexity confusion 27.73 (9.03) -.055 -.162

Market Experience 5.08 (1.34) -.686 .278 Source: this study

8.6.1. Examining the BPM-E (The Extensional Behavioural Perspective

Model)

This section will examine the patterns that the affective, behavioural, confusion and levels

of experience will indicate in the two market situations used in this study. The overall

reported levels for these variables will be reported and a comparison will be made in order

to examine is these follow the principles of the BPM. In order to compare the two

situations either one way Analysis of Variance (ANOVA) or t-test could be used.

Although t-test is the norm when only two means are to be compared, this study used

ANOVA as it is said to have more advantages in terms of helping to protect against Type


235

1 and Type 2 errors (Pallant, 2010). In this case an F-ratio is calculated. A significant F

test indicates that the null hypothesis (that there are no differences between the groups)

can be rejected. The following table indicates the means for all variables in this study per

situation and the ANOVA F test along with its significance.

In order to test for the assumption of the homogeneity of variances (Tabachnick & Fidell,

2007, p 85), tests like Levene’s test of homogeneity of variance have been introduced.

Levene’s test is provided by all ANOVA-like statistical processes. In this case the

preference would be for a significance level of greater than .05. In case this is not so, a

more stringent significance level should be used to judge the results of ANOVA. In case

Levene’s test sig is less than the .05 level a significance of .025 should be used for

moderate violation and of .01 for severe violation when evaluating the results of ANOVA-

Pallant, 2010. In case t-tests are used SPSS provides an output of both the equal and

unequal variance method and the correct level should be used to guide analysis.

Table 8.11 Means and ANOVA results for the two situations (N=260 for each

market)

Context Pleasure Arousal Domina

nce Approach

Avoida

nce A_A Similarity Complexity

Market

Exp/nce

Grocery (a) 35.70

(9.1)

23.70

(5.6)

26.48

(5.1)

11.22

(3.2)

9.30

(4.4)

1.88

(6.5)

16.68

(5.2)

26.33

(8.17) 5.32

PC/Laptop (b) 39.88

(8.7)

26.62

(5.1)

25.87

(5.6)

12.80

(3.3)

7.90

(3.8)

4.87

(5.8)

17.17

(5.7) 29.13 (9.6) 4.80

Anova Results

(Difference

between the

markets)

F(1,519)

=

28.430**

F(1,519)

=

38.857**

F(1,519)

=

1.253

F(1,519)

=

30.010**

F(1,519)

=

15.831**

F(1,519)

=

30.337**

F(1,519)

=

1.009

F(1, 159)

=

12.807**

F(1,519)

=

18.295**

**Difference significant at the 0.01 level (2-tailed)

*Difference significant at the 0.05 level (2-tailed)

Source: this study

The table indicates that a comparison of different markets in accordance with the

principles of the BPM is possible even when the situations are not part of the clear

boundaries of the operant classes and categories introduced by the original model.


236

Grocery shopping indicates significantly less utilitarian and informational reinforcement

than high technology buying. The levels of dominance reported differ with the high

technology market indicating lower levels of dominance however the difference is not

significant (most possibly a result of the two markets being part of today’s open retail

environments). This leads us to conclude that high technology buying should have more

approach and aminusa but less avoidance behaviour because it is overall characterised by

higher levels of reinforcement. This is in accordance with our data.

Levels of similarity confusion do not indicate significant differences between the two

markets while complexity is higher for the PC/Laptop market. It is interesting to note that

the levels of our behavioural variables (Approach-Avoidance) are not determined by

levels of punishment (levels of approach and avoidance might be influenced but not

determined by the difference of confusion between the two markets) but from the overall

levels of reinforcement. This is in accordance with theoretical arguments that say that it is

usually the level of reinforcement that defines the responses within a situation.

Overall, participants have reported that the grocery market has higher levels of

experience. The difference is significant although the markets do not differ as much as

expected.


237

8.6.2. Examining the BPM-I (The Intentional Behavioural Perspective

Model)

In order to perform this analysis similarity and complexity confusion levels had to be

reduced to levels in order to be able to conduct one-way ANOVA compare the groups of

confused and non confused consumers. Both a median and a mean split have been

considered following instructions from previous research (Ward & Banes, 2001). The

median split was implemented as dictated by the ‘visual binning’ process of SPSS.

Table 8.12 Mean and median for the two confusion variables

Mean Median

Similarity 16.93 17.00

Complexity 27.73 28.00

Source: this study

The results for similarity confusion are presented in table 8.13 below:

Table 8.13 Means and ANOVA results for similarity confusion (BPM-I)

Context Pleasure Arousal Dominance Approach Avoidance Aminusa

High similarity

Confusion 36.05 (9.3) 24.7 (5.7) 25.50 (5.2) 11.40 (3.2) 9.20 (4.0) 2.30 (6.0)

Low similarity

Confusion 39.4 (8.7) 25.5 (5.3) 26.07 (5.4) 12.50 (3.4) 8.00 (4.2) 4.40 (6.4)

Anova Results

(Difference

between the

intentional

situations)

F(1,519)=

18.456**

F(1,519)=

2.757ª

F (1,519)=

6.192*

F(1,519)=

11.030**

F(1,159=

10.174**

F(1,159)=

14.755**

Levene’s test F (1, 518)=

.217

F(1,518)=

2.661

F(1, 518)=

.852

F(1, 518)=

.496

F (1,518)=

.128

F (1,518)=

3.649

Levene’s significance .641 .103 .356 .482 .721 .057



ª .097

Source: this study

These results confirm the hypotheses (H1- H6) related to BPM-I. The situation which is

characterised as more open and indicates higher utilitarian and informational


238

reinforcement (low confusion as understood by consumers) will indicate higher approach

and aminusa but lower avoidance.

The results for complexity confusion are presented in table 8.14 below:

Table 8.14 Means and ANOVA results for complexity confusion (BPM-I)

Context Pleasure Arousal Dominance Approach Avoidance Aminusa

High complexity

Confusion 37.05 25.86 24.97 11.86 9.26 2.60

Low complexity

Confusion 38.50 24.57 27.11 12.10 8.07 4.04

Anova Results

(Difference

between the

intentional

situations)

F(1,519)=

2.843*

F(1,519)=

7.076**

F (1,519)=

21347**

F(1,519)=

.674

F(1,159)=

10.823**

F(1,159)=

6.637**

Levene’s test F(1,518)=

.394

F(1,518)=

2.805

F (1,518)=

.329

F(1,518)=

.260

F(1,158)=

.234

F(1,158)=

.631

Levene’s significance .530 .095 .567 .610 .629 .427



Source: this study

ANOVA results indicate that arousal (informational reinforcement) seems to be higher

(although non-significant) for consumers suffering from complexity confusion. This

finding might simply indicate that complexity confusion is at some level similar to the

measurement of information rate of environments and it might include elements of this

construct.

Overall, however, findings support the notion that confusion is an un-aroused situation,

most possibly indicating that the level of feedback on performance that arousal is capable

of measuring is the one connected to social recognition in the form of consumer social

status rather than the self- esteem caused from understanding and increased

experience/learning of situations.


239

8.6.3. The Relationships Between the Variables

This section will deal with the relationships of the variables of this study. Examining the

correlation matrix of the variables of this study is crucial because this step can give a first

indication of both the direction and the strength of the association between variables.

Pearson-product moment correlation (r) has been used (in line with the use of parametric

tests in this study). This coefficient is designed for continuous variables and this

assumption is met by the variables of this study. An initial indication on the relationships

among PAD, confusion and behavioural variables will be examined in this section.

Before getting into the examination of the direction and strength of correlations and in

order to test for the assumption of linearity (Hair et al., p. 75), preliminary analysis in the

form of scatterplots of all relationships was conducted. The point is that correlations

represent only the linear association between variables, non-linear effects will not be

represented in the correlation value. Thus, when non-linearity is not taken into

consideration the results might be an underestimation of the actual strength14

of the

relationship (Hair et al., p. 75). All scatterplots of the relationships between variables have

been examined and the ones indicating the relationships with the behavioural variables are

presented as an indication of the linear relationships found in this analysis in Appendix 6.

Due to the lack of correlation between similarity/ complexity confusion with arousal (refer

to table of correlations 8.16) the scatterplot of the relationship between arousal and

confusion is presented below.

14 Linearity is not as important for factor analysis. In case relationships are not linear FA can be used, with a

main issue that the results might not indicate the same levels of quality (DeVaus, 2002, p. 385).


240

Figure 8.2 Scatterplots of the relationship between arousal and confusion

Similarity confusion

Arousal

Complexity confusion

Arousal

Source: this study

The scatterplots indicate that curvilinearity is not an issue in this study. Some issues with

heteroskedasticity are evident in the relationship between similarity/ complexity confusion

and the behavioural variables. These indications will be considered in the different

analysis by applying the rules developed for dealing with this issue. When data are

grouped, meaning in ANOVA like applications, more stringent levels for accepting the

results, in light of the Levene’s test of homogeneity of variance in all cases, will be

applied (Tabachnick & Fidell, 2007, p. 86). In view of the fact that heterskedasticity is not

a problem for the interpretation of results because the results of any analysis are mainly

weakened but not invalidated due to this issue, this study has not performed the solution

of transformations. The problem with transformations is that the analysis is then only

limited to the transformed data and cannot be easily extended to the actual social


241

phenomena. Interpretations should be taken with extreme caution. Prior to the execution

of multiple regression the residual plots will also be examined and heteroskedasticity will

be taken into consideration. Following instructions by many expert academics and

statisticians (Allison, 1999; Tabachnick & Fidell, 2007) overall regression estimation is

not de-validated by heteroskedasticity.

Following the data examination for the linearity of the relationships, the actual

examination of pearson correlation coefficient values can be conducted. The value of the

Pearson correlation ranges from -1.00 to 1.00. 1.00 indicates a perfect positive correlation

and -1.00 a perfect negative. The value of r is the value that determines the strength of the

relationship. Several authors propose different interpretations of this correlation

coefficient. Pallant (2010, p. 126) is based on Cohen (1988) and provides the following

table to indicate the interpretation of the value of the correlation.

Table 8.15 Interpretation of pearson correlation (r)

r= .10 to .29 or r=-.10 to -.29 Small

r= .30 to .49 or r=-.30 to -.49 Medium

R=.50 to 1.0 or r=-.50 to -1.0 Large

Source: Cohen, 1988 as in Pallant, 2010.

Following this estimation the subsequent table (8.16) presents the findings of the

correlations between the variables:


242

Table 8.16 Correlation coefficients of this study (N=520)

Variable Mean SD Cronbach

Alpha Correlation Coefficients

P A D AP AV A_A SC CC Ex

Pleasure 37.79 (9.18) .923 1

Arousal 25.16 (5.5) .661 .433** 1

Dominance 26.17 (5.4) .731 .354** .113** 1

Approach 11.99 (3.3) .633 .641** .339** .256** 1

Avoidance 8.61 (4.16) .727 -.496** -.252** -.170** -.434** 1

AP_AV 3.38 (6.3) - .660** .342** .245** -.808** -.882** 1

Similarity

Confusion 16.93 (5.5) .916 -.177** -.066 -.204** -.189** .185** -.220** 1

Complexity

Confusion 27.73 (9.03) .903 -.155** .088 -.287** -.115** .217** -.202** .680**

1

Experience 5.08 (1.34)

1 item only

(overall,

macro

experience

with a

market)

.296** -.021 .340** .321** -.121** .247** -.244** -.364** 1

**Correlation significant at the 0.01 level (2-tailed)

*Correlation significant at the 0.05 level (2-tailed)

Source: this study


243

Table 8.16 on the correlations is overall in accordance with the expectations of the

Mehrabian and Russell (1974) model, the BPM and the expectations of this study. The

level of correlation between similarity and complexity confusion is an issue if variables

are to be entered in an analysis (especially regression) as predictors together

(multicollinearity will be discussed as an assumption of regression- see section 8.6.5).

Similar strong correlations have been reported in past research for variables underlying

the same construct (e.g. Zimet, 1988). The fact that EFA indicated two different

constructs with high correlation point to the fact that these constructs are dissimilar but

underlie the same theoretical idea.

The inter-correlations among pleasure and arousal, and pleasure and dominance are not

sufficiently high to breach Mehrabian and Russell’s assumption of the orthogonality of

these variables, though the values found could raise some concerns about

multicollinearity. This issue will be examined in the regression analysis based on the

values of VIF and tolerance.

A theoretically interesting finding in this table is the relationship between similarity-

complexity confusion and arousal. In both cases a significant relationship is lacking.

Similarity however has a negative relationship, while complexity a positive. The

theoretical implications of this finding for future research will be discussed in the

concluding chapter (chapter 10).

Two additional interesting correlations will be presented in the form of table 8.17 below:

Table 8.17 Correlation coefficients between kinds of confusion in the two

markets

Similarity

(High technology market)

Complexity


Similarity (Grocery) .323**

Complexity (Grocery) .394** **Difference significant at the 0.01 level (2-tailed)

Source: this study


244

These correlation coefficients indicate the correlation between similarity confusion and

complexity confusion in the two markets. These are not important for the structure of this

analysis per se. However, there might be important theoretical implications pointing to a

better treatment of confusion in future research. The main question would be whether a

contextual treatment of confusion is really worthwhile. These theoretical implications will

be discussed in more detail in the final discussion chapter (chapter 10).

8.6.4. The Effect of Consumer Confusion

Before examining the relationship of confusion with the emotional and behavioural

variables per market/ situation, the following table summarises the relevant correlation

coefficients per market:

Table 8.18 Correlation coefficients per situation

Similarity confusion Complexity confusion

Grocery PC/Laptop Grocery PC/Laptop

P -.137** -.247** -.082 -.305**

A -.111 -.051 -.002 .097

D -.105* -.284** -.156** -.382**

AP -.174** -.234** -.047 -.251**

AV .174** .220** .261** .244**

Aminusa -.201** -.276** -.197** -.302** **Correlation significant at the 0.01 level (2-tailed)


Source: this study

Overall, the direction of the relationships is as theoretically expected for all relationships

examined. It is negative for Pleasure, Dominance, Approach, Aminusa and positive for

Avoidance. Arousal seems to be the unique variable that has a more complicated

relationship. Specifically, all correlation coefficients are non significant, however all

relationships are weak but negative with the exception of the relationship between

complexity and arousal in the high-technology market which is weak, non-significant but

positive.


245

It is also important to note that the stronger relationships are between confusion and

pleasure, confusion and dominance and from the behavioural variables confusion

and avoidance. Previous theoretical arguments support this finding by indicating that the

absence of confusion will not increase satisfaction and approach but its presence will

increase dissatisfaction and avoidance.

In order to test for a possible combined effect of the market/situation (mainly representing

the levels of experience of each market) and confusion on each emotional and behavioural

variables (representing the ideas of BPM) factorial ANOVA will be used. Factorial

ANOVA allows for the exploration of both the main and interaction effects (Kinnear &

Gray, 2000). The main effect is the effect of each variable on its own whereby the

influence of the other is disregarded. An interaction effect occurs when the effect of one

variable is not the same under all conditions of the other variable (Pallant, 2010; Brace et

al., 2012). Both main and interaction effects are significant, when p≤ .05. When

interaction effects are present the interpretation of the main effects should take these into

consideration. This approach will then be used for the examination and establishment of

hypotheses 7-12. In order to complete this step again 2 groups of similarity and

complexity confusion were used (based on a median split as explained previously) and the

calculations were based on these and the two market levels as a dummy variable

(1=Grocery market and 2=PC/Laptop). The General Linear Model process of SPSS was

used for this analysis.

For the factorial ANOVA applications the guidelines for acceptance of results have

become equally strict as any other ANOVA. As already explained above, in order to test

for the assumption of the homogeneity of variances, tests like Levene’s test of

homogeneity of variance have been introduced. Levene’s test is provided by all ANOVA

like statistical processes and a more stringent significance level should be used to judge


246

the results (if levene’s test is significant at the .05 then a significance of .025 should be

used for moderate violation and of .01 for severe violation to accept the ANOVA results.

This principle has guided the following analysis.

In addition, eta-squared values (Cohen, 1988 as in Pallant, 2010), which indicate the

effect size of the results, will be provided. Cohen's guidelines (1988) indicate that: 0.01 =

small, 0.06 = medium, 0.13 = large effect size.

Table 8.19 summarises the results for similarity and table 8.20 the results obtained for

complexity confusion.

Table 8.19 Factorial ANOVA results for similarity confusion

Similarity

Confusion F (1, 519) P< Eta Squared

Pleasure

S. Confusion 24.071 .000** .05

Markets 33.882 .000** .06

Confusion*Markets .970 .325 -

Arousal

S. Confusion 5.126 .024* .01

Markets 41.511 .000** .08


Dominance

S. Confusion 5.751 .017* .03

Markets .904 .342 -

Confusion*Markets 1.979 .160 -

Approach

S. Confusion 15.255 .000** .04

Markets 34.155 .000** .06


Avoidance

S. Confusion 12.907 .000 .03

Markets 18.430 .000 .04


AP_AV

S. Confusion 19.833 .000** .04

Markets 35.297 .000** .07




Source: this study

Considering that eta-squared values typically range from .01 to .09 in the social sciences

(Cohen, 1988) and following Cohen's guidelines (1988 as in Pallant, 2010), that: 0.01 =


247

small, 0.06 = medium, 0.13 = large effect size, mainly small to medium significant main

effects were found in the study of similarity confusion. This indicates that the effect of

similarity confusion does not depend on the characteristics of each market and it possibly

signifies that the effect of similarity confusion is independent of the overall levels of

consumer experience as developed in specific choice situations. This finding has multiple

managerial consequences which will be discussed in the final chapter of this thesis

(chapter 10).

Table 8.20 Factorial ANOVA results for complexity confusion

Complexity

Confusion F (1,519) P<

Eta Squared

(variance

explained)

Pleasure

C. Confusion 6.672 .010* .02

Markets 30.291 .000** .06

Confusion*Markets 6.189 .013 .02

Arousal

C. Confusion 3.086 .080 .02

Markets 34.092 .000** .08


Dominance

C. Confusion 19.726 .000** .04

Markets .384 .536 -

Confusion*Markets 6.197 .013* .02

Approach

C. Confusion 2.890 .090 -

Markets 30.210 .000** .06

Confusion*Markets 5.360 .01* .015

Avoidance

C. Confusion 16.388 .000** .04

Markets 21.073 .000** .05


AP_AV

C. Confusion 12.795 .000** .03

Markets 35.253 .000** .07

Confusion*Markets 2.088 .149 -



Source: this study

Again mainly small to medium significant main effects were found in the study of

complexity confusion. Further to the above results and due to the fact that interaction

effects were found for the variables Pleasure, Dominance and Approach behaviour, the


248

following factorial matrices are presented for these three variables. The results indicate

that indeed complexity confusion has a stronger effect in the market with the overall lower

levels of experience (that of high technology market) when it comes to utilitarian

reinforcement, behaviour setting scope and approach behaviour. However, this effect is

obviously very limited in scope as the eta squares indicate. In addition, this finding does

not extend to the other two behavioural variables that of aminusa and especially

avoidance.

The cell means (factorial ANOVA matrices and plots) for the three variables (pleasure,

dominance, approach) are presented below:

Table 8.21Cell means for pleasure (complexity-situations)

Complexity

confusion

Plot

Grocery (habitual/everyday

market)

Low 35.7

High 35.6

High technology market

Low 42.04

High 38.03

Source: this study

Table 8.22 Cell means for dominance (complexity- situations)

Complexity

confusion

Plot


market)

Low 26.74

High 25.82


Low 27.6

High 24.4

Source: this study


249

Table 8.23Cell means for approach (complexity-situations)

Complexity

confusion

Plot


market)

Low 11.15

High 11.32


Low 13.40

High 12.14

Source: this study

8.6.5. The Degree of the Relationship between the Affective-Confusion and

Behavioural Variables

This section deals with the degree of the relationship between the affective and

behavioural variables. In addition to the typical PAD emotional reactions, this study will

add confusion to the model. Multiple regression analysis (main effects only) has been

used to assess the relationship between one DV and several IVs. The results of multiple

regression should not be confused with those of correlation, as correlation intents to assess

the relationship while regression is used for prediction (Tabachnick & Fidell, 2007, p.

117)

The first task before starting a multiple regression and more accurately described as the

assumptions that need to be met are the efficiency of the sample size to perform

regression, levels of multicollinearity and levels of lineariy and heteroskedasticity

(Tabachnick & Fidell, 2007, p. 123).

a. Sample Size

Regarding sample size, there are different guidelines concerning the number of cases

required for multiple regression. It has been said (Stevens, 1996, cited in Pallant, 2010, p.

148) that for social science research, ‘about 15 subjects per predictor are needed for a

reliable equation‘. Tabachnick and Fidell (2007, p. 123) state a formula for calculating


250

sample size whereby ‘N>50 +8m, m is the number of independent variables‘. The current

research has four independent variables; therefore, N should be more than 60 cases

(N>60) or 82 cases (N>82). It can be concluded that the sample size of this study clearly

satisfies the sample size assumptions required for multiple regression analysis.

b. Multicollinearity

Multicollinearity is a problem with multivariate techniques of data analysis that occurs

when independent variables are too highly correlated and the lack of multicollinearity is a

crucial assumption to meet in regression. The effect of a possible multicollinearity

problem would have been that small data changes or arithmetic errors could be translated

into very large changes or errors in the regression analysis. The issue with this is that the

calculation of the regression coefficients requires the inversion of the matrix of

correlations among the independent variables, and this cannot be made when

multicollinearity is present (Tabachnick & Fidell, 2007, p. 124). A rule of thumb that

signifies problems indicating further investigation is the mere fact that independent

variables are highly correlated (usually above 0.7). In addition, SPSS performs a relevant

analysis of ‘collinearity diagnostics’. The two statistical indexes used as the result of this

analysis are VIF (Variable Inflation Matrix) indicates the inflation of the variance of the

coefficient of regression as a consequence of the correlation between independent

variables and tolerance (1-R²; R² is the squared multiple correlation). VIF should not

exceed 5 and tolerance should not fall below 0.1 in order to indicate that a regression does

not suffer from multicollinearity. Additional, examinations would include the Collinearity

Diagnostics table (provided by SPSS as a result of regression); there the Condition Index

should not be over 30 and at the variance proportions column no variables should have

variance proportions more than 0.50. Multicollinearity was not an issue for the regressions

of this study (VIF and tolerance statistics are presented in the regression tables below).


251

c. Linearity and Homoskedasticity- Visual Examination of Residual

Linearity refers to the relationship between two variables (Hair et al., 1998, p. 75) and

homoskedasticity is an assumption related to the dependence relationships between

variables (Hair et al., p. 73). Specifically, linearity implies that there are linear

relationships between the variables as already described above. The most common way to

access it has been to examine scatterplots of the variables and to examine any nonlinear

pattern to the data. This has been done as described before in the Pearson’s correlation

matrix. No curvilinearity has been detected (see section 8.6.3 this analysis).

Homoskedasticity at the other end indicates whether the dependent variables signify equal

levels of variance across the range of independent variables (Hair et al., 1998, p. 73-75).

The existence of homoskedasticity is not fatal for the analysis of the data it is just that the

predictability is better if heteroskedasticity is accounted for. Specifically, Tabachnick and

Fidell (2007, p. 85 and p. 127) ascertain that heteroscedasticity in multiple regression does

not invalidate the results as much as weakens them. Allison (1999, p. 128) further argues

that heteroskedasticity is worth checking but it has to be pretty severe before it leads to

serious bias in the standard errors. The conclusion is: ‘Although it is certainly worth

checking, I wouldn’t get overly anxious about it.’

Anyhow, following the assessment of linearity this study has found it relevant to examine

for the assumption of homoskedasticity before running the regressions. Results can then

be interpreted with more certainty. This analysis has been conducted through the

inspection of the analysis of the regression residuals (Tabachnick & Fidell, 2007, p. 78).

Problems with regression are generally more easily judged when plotting the residuals

rather than the original data. Two types of residual plots have been examined following

instructions by Pallant (2010), the normal probability plot of the regression standardised

residuals and scatterplot of standardised residuals. Both residual plots are given along


252

with the results of the regression. The line of normal probability plot should lie in a fairly

straight diagonal line and in the second plot the residuals should be indicating a pattern of

being roughly rectangularly distributed (Pallant, 2010, p. 156). The plots for testing the

assumption of homoskedasticity and linearity together for the regression analysis are

provided in the next pages.


253

Figure 8.3 Normal P-P plots of regression standardised residual for

Approach

Figure 8.4 Scatterplots of standardised residuals for Approach

PAD+Similarity

PAD+Complexity

Source: this study


254


Avoidance

Figure 8.6 Scatterplots of standardised residual for Avoidance

PAD+Similarity

PAD+Complexity

Source: this study


255


AMINUSA

Figure 8.8 Scatterplots of standardised residual for AMINUSA

PAD+Similarity

PAD+Complexity

Source: this study

The inspection of the plots indicates that multiple regression can be performed with a high

degree of confidence for approach and aminusa behaviour. For avoidance behaviour the

line of the P-P plots is not a perfect straight line and there are some concerns for the level


256

of multivariate normality. These levels of abnormality have been however decided to be

tolerated by this study. The main rational for that is based on all previous assertions that

this problem has to be rather severe in order to have an effect on regression. Tabachnick

and Fidell (2007, p. 85 and p. 127) ascertain that heteroscedasticity in multiple regression

does not invalidate the results as much as weakens them. On these grounds multiple

regression was used for the prediction of all behavioural variables.

d. Performing the Regression

In regression there are two values reported indicating the overall fit of the model— R2 and

adjusted R2. As more independent variables are added to the regression model, unadjusted

R2

will generally increase. This will occur even when the additional variables do little to

help explain the dependent variable. Adjusted R2 is used to compensate for the addition of

variables to the model, because adjusted R2 is corrected for the number of independent

variables of the model (Brace et al., 2006).

Reporting the unadjusted R2

is acceptable but the adjusted R2 is the requirement when

there are multiple models presented with varying numbers of independent variables.

Based on these reasons the adjusted R2

will be reported in this section.

In addition, it is essential to state that by adding one or more independent variables to a

multiple regression model, two things can potentially occur as a result:

1. Change the overall R-square value

2. Change the contribution of the other independent variables

Thus, the comparison of two or more multiple regression models where variables have

been added is not efficient by just comparing the change in adjusted R- square value itself.

Adding a new independent variable might improve the adjusted R-square (at least a little),

but it is an optimisation of sorts that may mean reducing or increasing the contribution of


257

another variable. Thus the R2

change, stepwise regression SPPS process will be used and

will be reported specifically for consumer confusion which is the added element in the

model.

Model 1 is the model with the 3 PAD variables only and models 2 are the PAD variables

with the addition of either the similarity or complexity confusion. When both similarity

and complexity are added to the model, there are elements of multi-collinearity. This step

was not implemented as it was not seen as essential or imperative for the requirements of

this research.


258

Table 8.24 Regression for Approach Behaviour, N=520

Source: this study

15 Change attributed specifically to the addition of confusion.

Model F ( ) P<

Adjusted

R Square

R square

change15

P<

Toler

ance VIF

Model 1

Approach=

F(3,519)

=

123.083

.000 .414

Pleasure+ .59 .000 .718 1.393

Arousal+ .12 .01 .810 1.239

Dominance .037 .308 .873 1.146

Model 2 Approach=

F(4,519)

= 94.166 .000 .418

Pleasure+ .58 .000 .710 1.409

Arousal+ .12 .01 .810 1.234

Dominance+ .025 .497 .852 1.173

Similarity

Confusion .005 -.075 .03 .946 1.057

Model 2 Approach=

F(4,519)

= 92.291 .000 .413

Pleasure+

.59 .000 .706 1.416

Arousal+

.14 .001 .788 1.270

Dominance+

.031 .402 .820 1.220

Complexity

confusion .000 -.022 .544 .888 1.126


259

Table 8.25 Regression for Avoidance Behaviour, N=520

Source: this study

Model F ( ) P<

Adjuste

d R

Square

R

square

change P< Tolerance VIF

Model 1 Avoidance=

F(3, 516)=

56.571 .000 .242

Pleasure+ -.48 .000 .718 1.393

Arousal+ -.06 .278 .810 1.239

Dominance .004 .913 .873 1.146

Model 2 Avoidance=

F(4, 515)=

44.685 .000 .252

Pleasure+ -.46 .000

Arousal+ -.06 .278

Dominance+ .02 .606

Similarity

Confusion .010 .104 .008

Model 2 Avoidance=

F (4, 515)=

47.738 .000 .266

Pleasure+

-.45 .000 .709 1.411

Arousal+

-.07 .084 .788 1.268

Dominance+

.047 .269 .819 1.221

Complexity

confusion .025 .16 .000 .892 1.120


260

Table 8.26 Regression for AMINUSA Behaviour, N=520

Source: this study

Hypotheses 20-23 are examined above. Confusion has a rather small but significant

contribution to the overall model. This contribution is greater with avoidance and aminusa

and less with approach behaviour. Pleasure has the largest and strongest beta value (β) and

seems to be the most important element when considering behaviour in market/shopping

situations. Dominance has no contribution to the models.

Model F ( ) P<

Adjuste

d R

Square

R

square

change P< Tolerance VIF

Model 1 Aminusa=

F (3,516)=

135.151 .000 .437

Pleasure+ .624 .000 .718 1.393

Arousal+ .106 .03 .810 1.234

Dominance .016 .462 .873 1.146

Model 2 Aminusa=

F (4,515)=

105.703 .000 .447

Pleasure+ .611 .000 .710 1.409

Arousal+ .107 .002 .810 1.234

Dominance+ -.001 .978 .852 1.173

Similarity

Confusion .011 -.071 .004 .946 1.057

Model 2 Aminusa=

F (4,515)=

106.232 .000 .449

Pleasure+

.609 .000 .706 1.416

Arousal+

.116 .001 .787 1.270

Dominance+

-.015 .686 .820 1.220

Complexity

confusion .013 -.091 .002 .888 1.126


261

8.6.6. The Pattern of the Relationship between the Affective/Confusion

Variables and Aminusa

Finally, in order to explore hypothesis 24 on the possible interaction between pleasure and

arousal again the General Linear Model will be used. Following instructions from Soriano

et al., (2013) in this case all pleasure, arousal, dominance and both confusion variables

had to be subdivided into low, medium and high levels. This procedure was necessary to

carry out the study, but it should be noted that this subdivision of mean scores inevitably

reduces the accuracy of estimated interactions among continuous variables. The Tukey

HSD test was used in order to test the differences between the groups and the visual

binding of SPSS was used to guide the choice of groups’ subdivisions. Specifically,

pleasure was reduced into three groups were low (6-34), medium (35-42) and high (43

and more). Arousal was reduced into low (5-23), medium (24-27) and high (28 and more),

dominance low (5-24), medium (25-28) and high (29 and more), similarity confusion low

(4-15), medium (15-20) and high (20 and more) and finally complexity confusion low (7-

25), medium (26-32) and high (33 and more). The results are presented in the tables

below:


262

Table 8.27 ANOVA results for Approach-Avoidance (AMINUSA)

PAD+Similarity.

F Sig.

Main effects

Pleasure 44.789 .000

Arousal 9.836 .000

Dominance 1.484 .228

Similarity 2.637 .053

Two-way interactions

Pleasure*Arousal 3.629 .006

Pleasure*Dominance .611 .655

Arousal*Dominance .419 .795

Pleasure*Similarity .630 .641

Arousal*Similarity 2.825 .025

Dominance*Similarity .196 .940

Three-way interactions

Pleasure*Arousal*Dominance .366 .938

Pleasure*Arousal*Similarity 1.251 .268

Pleasure*Dominance*Similarity 1.053 .396

Arousal*Dominance*Similarity .655 .731

Pleasure*Arousal*Dominance*Similarity .590 .850

Source: this study

The results indicate (in accordance with the regression results) significant main effect for

Pleasure, Arousal and Similarity confusion and also an interaction effect between Pleasure

and Arousal. The table of the cell means for this interaction is presented below.

Table 8.28 Pleasure and arousal interaction. Cell means for AMINUSA

Arousal

Pleasure Descriptive

statistics

Low Moderate High Total

Unpleasant

environment

M -2.87 .10 1.6 -1.1

SD 5.1 6.2 5 5.8

N 88 83 21 192

Neutral

environment

M 2.17 5.06 4.67 3.99

SD 3.9 4.4 4.2 4.3

N 53 57 52 162

Pleasant

environment

M 7.38 7.77 8.3 7.96

SD 4.3 5.6 4.9 4.6

N 39 40 87 166 Source: this study

The Tukey HSD test has indicated that the means score for pleasure and arousal are highly

significantly (p<.001) different between the three levels of means (low, medium, high).


263

Table 8.29 ANOVA results for Approach-Avoidance (AMINUSA)

PAD+Complexity

F Sig.

Main effects

Pleasure 35.417 .000

Arousal 11.377 .000

Dominance .624 .536

Complexity 6.851 .001

Two-way interactions

Pleasure*Arousal 1.857 .117

Pleasure*Dominance .405 .805

Arousal*Dominance .451 .772

Pleasure*Complexity 1.449 .217

Arousal*Complexity 1.899 .110

Dominance*Complexity .559 .693

Three-way interactions

Pleasure*Arousal*Dominance .628 .754

Pleasure*Arousal*Complexity .524 .839

Pleasure*Dominance*Complexity 1.805 .084

Arousal*Dominance*Complexity .860 .551

Pleasure*Arousal*Dominance*Complexity .715 .759

Source: this study

The table indicates only main significant effects for pleasure, arousal and complexity

confusion. No interaction effects have been identified in this model.

8.7. The Results of the Statistical Hypothesis Testing

The main objective of this study has been to propose a novel understanding for the

concept of consumer confusion based on the theoretical principles of behaviourism and

thus extend the applications and ways of treating the construct. Confusion has been

measured and conceptualised based on previous studies of consumer behaviour as

similarity and complexity confusion and both of these conceptualisations are in

accordance with the idea of anomy- which implies the lack of market rules that can guide

behaviour, making confusion a punishing aspect of markets. The main aim has been then

to indicate that either as an extensional or intentional construct confusion can have

implications for consumer behaviour and will determine along with the bifurcation of

reinforcement and behaviour setting scope behavioural responses. Results of multiple


264

regression (tables 8.24- 8.26) and ANOVA (tables 8.27 and 8.29) provide support for this

assumption. Other hypothesis based on the previous studies and theoretical propositions of

the BPM were developed to further describe the relationships expected among the

variables. Overall, support is provided for the patterns expected with behavioural

variables following the pattern of reinforcement (BPM-E) and contingencies indeed being

modified by the persons’ rule making (BPM-I).

Table 8.30 summarises the predictions set by research hypotheses (as in chapter 6) and

further describes the findings of this analysis.


265

Table 8.30 Results of statistical hypothesis testing

Similarity Complexity

H1: Overall, the range of confused consumers will indicate lower levels of

Approach behaviour than the range of non-confused consumers.

Supported Not supported

No sig. difference.

H2: Overall, the range of confused consumers will indicate higher levels of

Avoidance behaviour than the range of non-confused consumers. Supported Supported


Aminusa (approach-avoidance) than the range of non-confused consumers. Supported Supported


Pleasure than the range of non-confused consumers. Supported Supported

H5a: Overall, the range of confused consumers will indicate lower levels of

Arousal than the range of non-confused consumers.

Not supported

(The difference is as

expected but small).

Not supported

H5b: Overall, the range of confused consumers will indicate the same levels of

Arousal with the range of non-confused consumers. Supported Partially supported


Dominance than the range of non-confused consumers. Supported Supported

H7: The effect of confusion on Pleasure will be stronger for the market


Not supported Supported

H8: The effect of confusion on Arousal will be stronger for the market

characterised by overall lower levels of experience. Not supported Not supported

H9: The effect of confusion on Dominance will be stronger for the market

characterised by overall lower levels of experience. Not supported Supported

H10: The effect of confusion on Approach behaviour will be stronger for the

market characterised by overall lower levels of experience. Not supported Supported

H11: The effect of confusion on Avoidance will be stronger for the market

characterised by overall lower levels of experience. Not supported Not supported

H12: The effect of confusion on Aminusa (approach-avoidance) will be stronger

for the market characterised by overall lower levels of experience. Not supported Not supported


266

H13: The two markets are expected to differ in terms of utilitarian reinforcement

with the high technology market expected to have higher Pleasure than the grocery

market.

Supported

H14: The two markets are expected to differ in terms of informational

reinforcement with the high technology market expected to have higher Arousal

than the grocery market.

Supported

H15: The two markets are expected to differ in terms of dominance with the high

technology market expected to have lower Dominance than the grocery market. No significant difference identified between the markets.

H16: Possible differences are expected in the levels of Confusion between the two

markets of this study.

Supported for complexity confusion,

Same levels for similarity confusion.

H17: Approach will be higher in the market characterised by higher levels of

utilitarian and informational reinforcement (thus the high technology purchasing

situation).

Supported

H18: Avoidance will be higher in the market characterised by lower levels of

utilitarian and information reinforcement (thus the grocery market is expected to

have higher avoidance).

Supported

H19: Aminusa, the net difference between approach and avoidance will be higher

in the market characterised by higher levels of utilitarian and informational

reinforcement (thus the high technology purchasing situation).

Supported

H20: Affective variables of Pleasure, Arousal and Dominance will each have a

positive relationship with Approach. Confusion will have a negative relationship.

Supported Supported


negative relationship with Avoidance. Confusion will have a positive relationship.

Supported for P, A, SC. Not

supported for dominance.

Confirmed for P, A, CC. Not



positive relationship with Aminusa, the net difference between Approach and

Avoidance. Confusion will have a negative relationship.

Supported for P, A, SC. Not


Supported for P, A, CC. Not


H23: Aminusa (the net difference between Approach- Avoidance) will be

determined by the variables pleasure, arousal, dominance and confusion.

Supported for P/A/SC.

Not for dominance

Supported for P/A/ CC.

Not for dominance

H24: Two-way interactions can be identified between the affective variables

Pleasure and Arousal when examining their effect on Aminusa. Supported for the model of similarity confusion.

Source: this study


267

8.8. Conclusion

Chapter 8 deals with the exploratory/descriptive findings obtained from the empirical

survey. Data were analysed based on a variety of different types of analysis. Analysis was

based on statistical principles but also the norms established in previous relevant studies.

Different tests were conducted in order to describe both the sample’s characteristics and

the data, to establish reliability and validity and to test the hypotheses. Overall, the

analysis provides support for the majority of the hypothesis and expectations but not for

others as discussed and presented in the table above. An extended discussion of the

findings and the contributions of this study will follow in Chapter 9 (further analysis) and

10 (discussion, implications and future research).

CHAPTER 9- FURTHER ANALYSIS

268

9. FURTHER ANALYSIS

9.1. Introduction

Chapter 9 provides a discussion of the research findings in the context of some further

analysis, which mainly examines the comparison of these results with previous studies

and findings. The chapter is divided into several parts: (1) the first part is concerned with

an examination of whether consumer socio-demographic characteristics influence

perceptions of confusion. The results of this inquiry corroborate previous findings which

argue that socio-demographics are not a consistent segmentation predictor, (2) a

discussion and comparison of the scales’ reliability and elements of validity as applied in

this and previous studies will be discussed in part two and finally (3) the ways that the two

situations examined as part of this study’s BPM-E, could fit the operant classes and

categories of the BPM will be implemented. This analysis will be based on previously

reported levels of the PAD and behavioural variables. As part of this section the prospect

of organising previous findings and establishing a unique range for each contingency

category will be examined.

In these two last sections, the current findings are to be compared with some of the results

of previous applications of the original BPM in England, Venezuela and Wales (Foxall,

1997a; Foxall, 1997b; Foxall & Greenley, 1998; 1999; Foxall & Soriano, 2005).

9.2. Intentional Confusion and Consumer Socio-Demographic Characteristics

The marketing literature has attempted to use many variables in order to profile, target and

segment consumers. These include variables like geographics, cultural, personality and

importantly socio-demographic characteristics. Especially socio-demographic

characteristics have been used widely by companies and academic researchers for two

reasons: these are readily and easily available data and can be applied to segmentation


269

with relative ease (Myers, 1996). Despite this wide application, several previous studies,

ranging from environmental behaviour and levels of confusion (Diamantopoulos et al.,

2003; Walsh & Mitchell, 2005b), have identified a limited or to the best ambiguous value

of socio-demographic characteristics for segmenting and targeting consumers. Therefore

empirical findings on this issue can extend existing knowledge by pointing towards the

proper direction. That said if socio-demographics fail to be an adequate segmentation

basis then more complicated segmentation and targeting approaches ought to be applied in

the attempt to understand and explain consumer behaviour (Myers, 1996).

Based on the examination of confusion as an individual self-based rule, determined by the

experiences of an individual in the environment, it is then of value to extend our

understanding to additional issues like the effect of socio-demographic characteristics on

the levels of this rule-governed behaviour in each market.

Walsh and Mitchell (2005b) aimed to establish the socio-demographic characteristics of

consumer who find it difficult to choose. They described their research as both unique and

useful based on the segmentation possibilities of this approach. Specifically they tried to

investigate whether confusion is positively related to age (older consumers are more likely

to report higher levels of confusion) and negatively related to education (less educated

consumers might be more vulnerable to ambiguous environments). Using a sample of 264

consumers, drawn to represent the shopping public in a northern German city they

administered a questionnaire and tried to report differences based on their socio-

demographic characteristics. Their attempt failed to provide imposing evident supporting

the proposed relationships. They could identify a difference in the age group of 45-53 who

were found to perceive more confusion than those aged under 29, but that was the only

identified difference when it comes to age. In regard to education again only one


270

significant difference between participants having attended secondary school and those

with a university degree is reported.

The advantage of the present study compared to the above findings is that it has been

conducted in two contexts/markets rather than at the level of general confusion from the

market place (which can be described as a personality tendency), thus more distinct

differences might be determined. For this reason and in order to add to the complete

understanding of the phenomenon, analysis will be conducted in the contextual level of

the two markets rather than at the aggregate, overall level.

The findings of this study are very much in accordance with the previously mentioned

results and could not find impressive evident that socio-demographics are a useful

segmentation basis for categorising and treating consumer confusion. ANOVA was used

in order to determine differences between groups as described in table 9.1. Table 9.1

includes all findings for 1) age; 2) gender; 3) levels of education, and finally 4) the

differences between consumers who choose to shop online or in-store. The results are

presented below per market.


271

Table 9.1 ANOVA results for socio-demographic characteristics and levels of

confusion in the two markets

1. Age (for the purposes of

this ANOVA age was

transformed to a new

variable with 3 levels rather

than 6)

Levels Means N Anova-F Sig.

Grocery

Market

Complexity

confusion

1= 18-34

2=35-54

3= 55-65+

26.31

26.54

25.51

111

120

29

F (2, 259)=.182 .833

Similarity

confusion

1=18-34

2=35-54

3=55-65+

16.6

16.9

16.4

111

120

29

F(2, 259)=.088 .916

High

technology

market

Complexity

confusion

1= 18-34

2=35-54

3= 55-65+

28.4

29.5

30.4

111

120

29

F(2, 259)= .695 .500

Similarity

confusion

1=18-34

2=35-54

3=55-65+

16.7

17.2

18.5

111

120

29

F(2, 259)=

1.108

.332

2. Gender Levels Means N Anova-F Sig.

Grocery

Market

Complexity

confusion

Male

Female

27.2

25.5

123

136

F (1, 258)=

2.491

.116

Similarity

confusion

Male

Female

17.2

16.1

123

136

F(2, 258)=

2.863*

.04

High

technology

market

Complexity

confusion

Male

Female

26.9

31.2

123

136

F(1,

258)=13.289**

.000

Similarity

confusion

Male

Female

16.5

17.8

123

136

F(1, 258)=

3.179*

.02

3. Education16

(for the

purposes of this ANOVA

education was transformed

to a new variable with 2

levels)

Levels Means N Anova-F Sig.

Grocery

Market

Complexity

confusion

Lower

education

Higher

education

25.5

27.0

131

129

F(1, 253)=

2.243

.135

Similarity

confusion

Lower

education

Higher

education

16.2

17

131

129

F(1, 253)=

1.172

.280

16 GCSE/A levels-Vocational and technical school were coded 1/ Higher education and Postgraduate

degrees and other were coded 2.


272

High

technology

market

Complexity

confusion

Lower

education

Higher

education

29.3

29.2

131

129

F(1, 253)=.005 .945

Similarity

confusion

Lower

education

Higher

education

17.04

17.53

131

129

F(1, 253)=.486 .487

4. Shopping channel Levels Means N Anova-F Sig.

Grocery

Market

Complexity

confusion

Online

In-store

26.8

26.2

40

220

F(1, 254)= .155 .694

Similarity

confusion

Online

In-store

17.9

16.3

40

220

F(1, 254)=

2.723*

.043

High

technology

market

Complexity

confusion

Online

In-store

27.2

31.12

128

133

F(1, 258)=

11.158**

.001

Similarity

confusion

Online

In-store

16.4

17.9

128

133

F(1,258)=

4.328*

.038

Source: this study

Commencing on the findings, it is evident that levels of similarity and complexity

confusion in the high technology market increase with age, however according to Tukey’s

test none of these differences are significant. Remaining in the high technology market

differences are observed in gender with females reporting higher levels of complexity and

similarity confusion than males. Although the result is still not significant it is interesting

to note that in the grocery market consumers of lower education report lower levels of

complexity confusion. Although this study is against stereotyping, this result might

indicate that less educated consumers spend more time to choose their grocery products

(more discretionary time and less income) and are thus more familiarised with the

industry. Finally, consumers who shop in store for PC/Laptops report to be significantly

more confused than those shopping online. This is an interesting finding that might be

explained on multiple levels and which should concern retailers of high technology

products. It could be assumed that either consumers who shop online are doing so because

they are not confused and thus they feel capable to cope with this task uninstructed by


273

store staff. In that same logic applies the fact that consumers who feel confused, prefer to

shop in-store in order to get help and instructions from store staff. Finally, an alternative

explanation is that shopping for high technology products in store is more confusing than

online shopping. It is however a finding that can have multiple implications for retailers

and the exact cause of this finding should be examined further. Finally, the finding that

consumers who shop for groceries online suffer from bigger levels of similarity confusion

should concern food and grocery retailers and especially web-developers and marketers.

9.3. The Scale of Confusion

The subsequent tables 9.2 and 9.3 will compare the results of this study with the ones of

previous applications of the measurement of confusion.

Table 9.2 A comparison of the internal consistency of the confusion measures

Walsh & Mitchell

(2010) study

This study

Number

of items

Cronbach’s

alpha

Number of

items

Cronbach’s

alpha

Similarity 3 .59 4 .916

Overload 3 .70 7 .903

Ambiguity 4 .75

Source: Walsh et al., 2007; Walsh & Mitchell, 2010; this study.

Table 9.3 A comparison of the correlations of the confusion measures

Walsh & Mitchell (2010) study This study

Similarity Overload Ambiguity Similarity Complexity

Similarity Similarity

Overload .310 Complexity .680

Ambiguity .141 .441

Source: Walsh & Mitchell, 2010; this study.

For the more efficient comparison of the results, it has to be reminded that this study

applied the scale in a contextual manner; meaning that the measurement involved the

levels of confusion in specific situations.

A comparison of the results above indicate that similarity confusion has demonstrated

better levels of reliability in this study as all four items proposed for the measurement


274

loaded correctly and with a rather high reliability coefficient alpha. The next important

implication is that overload and ambiguity confusion have clearly loaded on one factor

which based on theoretical reasons has been conceptualised as complexity confusion.

These two factors have actually indicated the higher correlation of all three constructs

(.441) in previous research also. The theoretical implications of this finding will be

discussed in the discussion chapter (see section 10.3.2). Finally, similarity and complexity

confusion indicate a higher correlation (.68) in this study. This finding should be assessed

by future research in case both kinds of confusion are to be used as indicators in a

multivariate model (regression or SEM). This high level of correlation might be the cause

of multicollinearity concerns and should be noted.

9.4. The PAD and Behavioural Variables

This study has further used the measures of the PAD and AP-AV behaviour as developed

by Mehrabian & Russell (1974). The decision to use these scales was facilitated based on

previous research findings which have proven the reliability and validity of the scales

(Foxall, 1997b, Soriano & Foxall, 2006). Although some researchers have questioned the

use of these scales in consumer environments (e.g. Donovan et al., 1994), the conceptual

connection of this study with the study of the BPM drove the decision to use the specific

measurements. In accordance with these observations, it was important to re-test the

unidimensionality and reliability of the scales due to the nature of the present study, which

used different consumer situations in a new research context. Thus, exploratory factor

analysis was used to assess the unidimensionality of the measurement scales for the

present study via the Principal Component Analysis (PCA) technique. The following

tables present the findings of this study, compared with previous studies of the eight

contingency situations.


275

Table 9.4 A comparison of the internal consistency of the measures

English Sub-study Cardiff Study Venezuelan Sub-study

(1)

This study

Number

of items

Cronbach

’s alpha

Number

of items

Cronbac

h’s

alpha

Number

of items

Cronbach’

s alpha

Number

of items

Cronbach’

s alpha

Pleasure 6 .88 6 .93 6 .93 6 .92

Arousal 6 .82 6 .83 6 .75 5 .66

Dominance 6 .89 6 .85 6 .82 5 .73

Approach 3 - 3 .74 3 .77 3 .63

Avoidance 3 - 3 .64 3 .79 3 .73

Source: Foxall & Greenley, 1999; Foxall & Soriano, 2011; Foxall & Soriano, 2005; this study.

Table 9.5 A comparison of the correlations of the PAD measures

English Sub-study Cardiff Study This study

Pleasu

re

Arous

al

Domina

nce

Pleasu

re

Arous

al

Domina

nce

Pleasu

re

Arous

al

Domina

nce

Pleasure

Arousal .26** .50** .43**

Domina

nce .33** .28** .50** .38**

.35** .11**

Source: Foxall & Greenley, 1999; Foxall & Soriano, 2011; this study.

The present findings show that the items of the PAD scale are uni-dimensional, however

rather strongly associated with each other, which is not unusual when the scale is applied

in other settings as table 9.5 indicates. However multicollinearity was not a problem in

this study (and equally not in any of the previous studies). Factor analysis revealed that

one item of arousal (A6. Sleepy- Wide-awake) and one of dominance (D2. Awed-

Important) loaded equally on their corresponding factor and Pleasure and had to be

removed from the constructs. Thus although 6 items loaded for Pleasure, only 5 items

worked for arousal and dominance.

Regarding reliability, all of the scales indicated acceptable alpha coefficients (>.60).

Nevertheless previous research which included all diverse categories of the BPM

contingency matrix indicated higher values, especially considering arousal and approach

behaviour.


276

9.5. Placing This Study’s Situations in the Context of the BPM Operant

Classes and Contingency Categories

In comparison to the previous studies of the BPM, the present inquiry has utilised

different consumer situations; however, the same measurement scales have been used,

perceived as underlying the same theoretical perspective. Beyond the main contribution of

this thesis which is the extension and study of consumer confusion based on an alternative

theoretical framework and importantly an extension of the behavioural perspective model

to an intentional mode, some additional research problems have been dealt in this study.

These issues include the consumer feelings and behaviour when it comes to specific

descriptions of retail situations and whether the BPM is capable of providing an

explanatory framework for the findings when situations are not manipulated.

Previous studies have measured situations that were specifically chosen to vary in terms

of the main BPM variables- see Foxall, 1999a for an analysis of the ‘consensibility and

consensuality’ of the description of situations used in the exploration of the BPM. Thus

situations were chosen and established to vary in terms of utilitarian-informational and

openness-closeness levels in accordance with the contingency categories of the BPM.

Although such a treatment has been adequate in establishing the PAD scales as sufficient

measurements of the constructs of reinforcement and behaviour setting scope, further

exploration of the applicability of the model is sought with this study. Thus, discussion

can be fruitful in identifying similarities or differences among the previous studies and

possibly placing the present situations in the boundaries of the operant classes and

contingency categories of the original model.

Following a comparison of the two situations the following table can be presented on the

conceptual comparison of the two markets of this study.


277

Table 9.6 Conceptual comparison of the two retail situations (based on BPM-E)

Grocery High Technology

pleasure PLEASURE Utilitarian

reinforcement

arousal AROUSAL Informational

reinforcement

dominance dominance

Closeness/ Openness

PC/Laptop indicated

lower levels of

dominance

(difference not

significant).

approach APPROACH

Approach increases

with the total quantity

and quality of

reinforcement, thus

high technology

buying has been

expected to have

more approach.

AVOIDANCE avoidance

Avoidance decreases


and quality of

reinforcement, thus

high technology

buying has been

expected to have less

avoidance.

aminusa AMINUSA

Aminusa increases


and quality of

reinforcement, thus

high technology

buying has been

expected to have

higher approach.

similarity confusion similarity confusion No effect of levels

of confusion on the

comparison of the

levels of the

behavioural

variables.

complexity confusion COMPLEXITY

CONFUSION

Higher levels of

complexity confusion

for buying of high

technology.

Source: this study (capital letters indicate higher levels of the specific variable, small letters lower,

same way of writing indicates same levels)

Table 9.6 facilitates the conceptual comparison of the two markets measured in this study.

The table is sufficient to indicate the differences between the two markets it cannot


278

however describe their positioning relatively to the operant classes and categories of the

BPM. During the development of the theoretical framework (see chapter 6) it was subject

to discussion whether the two specific retail situations will be part of the same

contingency category. Grocery shopping has been consistently been used as a situation

representing ‘maintenance’ and ‘routine purchasing’ (CC7) and the possible position of

the high technology buying situation was questioned and is worthy of investigation. The

high technology market was perceived as possibly sharing some qualities with the

categories of the ‘hedonism’ operant class. This was however described simultaneously as

a possibly ambiguous fact due to the nature of PC/Laptops. Although PC/Laptop use can

be connected with indulgence, these are nowadays very much work-related products.

In the boundaries of an attempt to provide an understanding of this issue, a table that

compiles the possible range and means that could meaningfully describe or define each

one of the eight contingency categories of the BPM has been constructed. The table has

summarised the findings of 8 studies conducted in England, Venezuela and Wales, based

on both the range of variables’ means (higher and lower means found in the relevant

studies) and the overall mean of the studies. It should be reminded that the situations used

in some studies differed, but in every case the main relationships as defined by the BPM

were identified from the corresponding research (higher pleasure for situations maintained

by higher utilitarian reinforcement, higher arousal for situations maintained by higher

informational reinforcement and higher dominance characterising open situations).


279

Table 9.7 Range and Means of PAD and AMINUSA for the eight published

studies of the BPM

Co

nti

ng

ency

Cat

egory

Range of

Pleasure (higher and lower

means reported

in relevant

studies)

Mean of

Pleasure (8 studies)

Range of

Arousal

(higher and

lower means

reported in

relevant studies)

Mean of

Arousal

(8

studies)

Range of

Dominance

(higher and lower

means reported in

relevant studies)

Mean of

Domina

nce

(8

studies)

Range of

Aminusa

(higher and lower

means reported in

relevant studies)

Mean of

Aminusa

(8 studies)

1 43.33-51.57 47.39 38.33-45.09 41.49 35.72-46.00 39.86 5.49-13.78 9.49

2 36.93-45.73 43.30 37.83-43.52 40.20 26.37-33.74 29.40 1.87-12.12 7.22

3 37.96-47.20 44.04 27.17-35.24 31.69 34.00-40.89 36.34 -0.07-8.44 5.96

4 27.51-40.75 34.51 22.89-30.64 26.78 18.96-29.66 25.73 -7.24-1.22 -3.88

5 33.99-44.39 37.05 34.30-41.92 37.93 33.33-40.38 36.37 -4.06-0.85 -1.50

6 26.89-41.33 33.82 30.35-38.73 35.67 27.67-32.44 29.44 -6.72-1.98 -2.18

7 27.96-37.22 32.53 25.53-31.92 27.56 33.19-38.82 35.41 -4.19-4.71 -0.47

8 21.73-38.97 25.31 22.92-31.04 27.14 19.52-28.71 24.77 -6.58-0.21 -4.13

Source: Compilation of the eight studies published in the following journal articles and book

chapter: Foxall, 1997b; Foxall, 1997c; Foxall & Greenley, 1999; Foxall & Soriano, 2005; Foxall

& Soriano, 2011; Foxall et al., 2012.

The main issue is that the attempt to define a specific range that could uniquely

characterise each category has failed. There are too many overlaps both at the level of

range and the levels of means which make the task of defining specific numeric

boundaries for the categories’ unattainable. It is a fact that dissimilar situations have been

used to describe the categories in the past (for example hedonism has been described by

situations ranging from been at the cinema to being at a party and fulfilment with

situations ranging from driving an expensive car to enjoying a luxurious holiday at an

exotic island). The comparison among the categories of each study is taking place among

the specific descriptions of situations of every study. However based on this table (table

9.7) and the findings of this study (table 9.8), it will be attempted to compare the two

situations used in this study with the numbers of previous studies.


280

Table 9.8 Means for the two situations examined in this study. Grocery and

high technology products (PC/Laptop) buying

Context Pleasure Arousal Dominance Approach Avoidance A_A

Grocery (a) 35.70 (9.1) 23.70 (5.6) 26.48 (5.1) 11.22

(3.2) 9.30 (4.4) 1.88 (6.5)

PC/Laptops

(b) 39.88 (8.7) 26.62 (5.1) 25.87 (5.6)

12.80

(3.3) 7.90 (3.8) 4.87 (5.8)

Anova Results

(Difference

between the

markets)

F(1,519)=

28.430**

F(1,519)=

38.857**

F(1, 519)=

1.253

F(1,519)=

30.010**

F(1,519)=

15.831**

F(1, 519)=

30.337**

Source: this study

In order to facilitate comparisons and in light of the fact that five out of the six items of

arousal and dominance formed the constructs in this study, the table has been transformed

by dividing all values with 6 (the number of variables that loaded correctly in all of the

previous BPM studies).


281

Table 9.9 Range and Means of PAD and AMINUSA for the eight published

studies of the BPM (all means divided by 6 so as to facilitate comparison)

Co

nti

ng

ency

Cat

egory

Range of

Pleasure (higher and lower

means reported

in relevant

studies)

Mean of

Pleasure

(8 studies)

Range of

Arousal

(higher and

lower means

reported in

relevant studies)

Mean of

Arousal

(8 studies)

Range of

Dominance

(higher and

lower means

reported in

relevant

studies)

Mean of

Dominance

(8 studies)

Range of

Aminusa

(higher and lower

means reported in

relevant studies)

Mean of

Aminusa

(8 studies)

1 7.2- 8.6 7.9 6.4-7.5 6.9 5.9-7.6 6.7 5.5-13.8 9.49

2 6.2- 7.6 7.2 6.3-7.2 6.7 4.4-5.6 4.9 1.8-12.1 7.22

3 6.3-7.8 7.3 4.5-5.8 5.3 5.6-6.8 6.0 -0.07-8.5 5.96

4 4.6-6.8 5.7 3.8-5.1 4.5 3.2-4.9 4.3 -7.2-1.2 -3.88

5 5.7-7.4 6.1 5.7-6.9 6.3 5.5-6.7 6.0 -4.06-0.85 -1.50

6 4.5-6.9 6.4 5.0-6.5 5.9 4.6-5.4 4.9 -6.7-1.98 -2.18

7 4.7-6.2 5.4 4.2-5.3 4.6 5.5-6.4 5.9 -4.2-4.7 -0.47

8 3.6-6.5 4.2 3.8-5.2 4.5 3.2-4.8 4.1 -6.6-0.21 -4.13

Source: Compilation of the eight studies published in the following journal articles and book

chapter: Foxall, 1997b; Foxall, 1997c; Foxall & Greenley, 1999; Foxall &Soriano, 2005; Foxall &

Soriano, 2011; Foxall et al., 2012.

Table 9.10 Means for the two situations examined in this study. Grocery and

high technology products (PC/Laptop) buying.

Context Pleasure Arousal Dominance Approach Avoidance A_A

Grocery

(a) 5.95 4.74 5.30 3.74 3.10 1.88 (6.5)

PC/Laptop

s (b) 6.64 5.32 5.17 4.26 2.63 4.87 (5.8)

Anova

Results

(Difference

between

the

markets)

F(1,519)

=

28.430**

F(1,519)

=

38.857**

F(1,519)=

1.253

F(1,519)=

30.010**

F(1,519)=

15.831**

F(1,519)=

30.337**

Source: this study

Starting with the grocery market, pleasure and arousal levels are very much within the

boundaries of category 7 ‘routine purchasing’, although pleasure levels reported in this

study are to the higher end of the range for this category. Dominance is however to the


282

lower end of this category and below the range reported in previous studies. Aminusa (the

net difference between approach and avoidance) is again within the range, although to the

higher end of the range.

Interestingly the mean of pleasure in the high technology market is above the mean and

range of situations usually positioned and conceptualised as measuring category 7 ‘routine

purchasing’. These levels of pleasure indicate that it could be positioned within the

hedonism operant class (it is within the range measured for categories 3 and 4 which

represents hedonism). The levels of arousal position high technology market closer to the

contingency category 3 ‘popular entertainment’ as the mean of arousal is closer to the

mean of this category and it also higher than ‘routine purchasing’. In terms of dominance

matters are also complicated as the mean of 5.17 for dominance for the high technology

market seems to position this market towards the categories of the matrix characterised as

closed. The mean for aminusa indicates a similar pattern. It is rather higher than ‘routine

purchasing’ and could be more relevant with category 3 ‘popular entertainment’.

In conclusion, it is easy to compare the levels of reinforcement and behaviour setting

scope of the two markets to each other and the results indicate the differences between the

two situations. Specifically, high technology market offers higher utilitarian and

informational reinforcement and these markets seem to be characterised by similar levels

of closeness and openness, however finding their exact positioning in the situations

established by the BPM seems a difficult task.


283

9.6. Conclusion

This chapter has provided some further analysis of the data of this study. First of all,

consumer socio-demographic information was examined in order to establish possible

differences in the way different categories of consumers perceive confusion in the two

retail markets. Further to that, a juxtaposition of the present findings with previous studies

of the BPM has been conducted. Based on this comparison high technology market was

found to only partially belong to the contingency category of ‘routine purchasing’ as it

has rather higher levels of utilitarian reinforcement and rather lower levels of dominance

than expected from the range determined from previous studies. It is applicable then to

argue that there are situations beyond the ones proposed by the contingency matrix of the

BPM which might have partial characteristics from one of the categories of the model and

partial from another. For example, high-technology shopping cannot be positioned within

the category of ‘routine purchasing’ per se but it might have elements of classes that are

characterised by higher levels of utilitarian and informational reinforcement, like those of

hedonism. Evidently, even in these cases the PAD and A_A measurement are able to

provide consistent results.

CHAPTER 10- DISCUSSION

284

10. DISCUSSION, IMPLICATIONS AND FUTURE

RESEARCH

10.1. Introduction

This thesis will conclude with the discussion of the research findings. Apart from the

arguments on the attainment of the main research objective which is the inclusion of rule-

governed behaviour and intentionality in the BPM, other areas of interest like the

implications for theory refinement and the study of confusion, the Behavioural

Perspective Model (Foxall, 1990) and the Mehrabian and Russell (1974) approach,

managerial implications, directions for future research and reflections on the process will

be offered in this section. The theoretical and methodological contributions of this study

will be discussed adopting both a chapter by chapter and a holistic approach.

10.2. Research Overview

This quest has described the BPM as an alternative model of consumer behaviour which is

based on the principles of operant conditioning. The model dictates that the rate of

recurrence of economic behaviour depends on the outcomes that similar behaviour has

had in the past (Foxall, 1992b). The principles of operant conditioning differ from those of

cognitive psychology, mainly because operant conditioning avoids the use of any

cognitive terms which implicate mental and internal to the individual processes. Without

implying that one approach is superior to the other but arguing that there are limitations to

both and that both can help knowledge to grow from their own perspective, Foxall (1990

and all subsequent research) developed the BPM which is based on a bifurcation of

situational consequences informed by consumer behaviour and consumer situation.


285

In what has been described as an example of ‘academic honesty’ (Oliveira-Castro, 2013,

p. 130), the problematic of the specific model has been identified, which mainly

concentrates on the hindering of the explanation of some aspects of behaviour.

Specifically, aspects like the continuity of behaviour, the personal level and the

delimitation of behavioural explanation (Foxall, 2004; Foxall, 2007b) seem to require an

alternative non-behavioural treatment and elucidation.

One of the suggestions to solve this issue involves the treatment of other kinds of

psychological concepts, found in ordinary language, that have been avoided by

behaviouristic approaches. For example, it has been suggested that the model should

consider dispositional concepts, in general, which include, in addition to propositional

attitudes, abilities, propensities, and personal emotions, or even personality traits (Foxall,

2007b; Oliveira-Castro, 2013). Considering that dispositional concepts and intentional

idioms in general describe, imprecisely, what individuals have done and predict what they

are likely to do under certain situations, they are good candidates to be included in the

description of consumers’ learning history (Oliveira-Castro, 2013). This is the reason that

Foxall (2007, p. 43) argues that intentional ascription is the result of the intersection of the

individual and the experiences in the specific situation.

Further to that, rules that derive either from others or the self (refer to rule-governed

behaviour) are good candidates to explore the two alternative BPMs proposed, the BPM-E

and the BPM-I (Foxall, 2013). The first model (BPM-E) advocates the use of an

extensional language (the language of objective facts) in terms of descriptions of ‘brute

facts’ and the second (BPM-I) the use of an intentional language where internal to the

individual beliefs and dispositional entities are allowed and facilitate the exploration of

behaviour. Rules are such aspects of human behaviour that can allow treatment at both

levels (Searle as in Foxall, 2013).


286

An alternative concept proposed and explored in this thesis is the idea that consumer

confusion can be treated as a self-based rule and specifically as a case of track. Tracks as

cases of rules are usually responses to the state of affairs- usually environmental affairs

(Zettle & Hayes, 1982). In that sense environmental arrangements are dictated by the

marketplace and confusion is a result of the state of affairs and environmental

arrangements (Glenn, 1987; Foxall, 2013). Taking this argument a step further the concept

of anomy (McClosky & Schaar, 1965), in terms of the lack of clear or concise rules that

can guide consumer behaviour, has been introduced in order to elucidate the concept

further. This treatment of confusion is a novel proposal for the construct which in the past

has been mainly treated and examined within the boundaries of the eminent cognition-

emotion debate (Zajonc, 1980; Lazarus, 1984).

Based on this theoretical framework (chapter 6) two models have been built and

examined. The first examines the BPM-E and treats confusion as an overall aversive

consequence of being and shopping in retail settings. In this case the levels of confusion

reported for each consumer situation have been treated at an overall level and describe the

levels of similarity and complexity confusion that being at different situations entail.

Concurrently, levels of utilitarian, informational and closeness- openness of the settings

have been described in accordance with the principles of the BPM. It is relevant to argue

that according to the principles of the BPM these variables should be perceived as

orthogonal, in the sense that any level of any variable could be accompanied by any levels

of the other (Foxall & Soriano, 2005; Foxall et al., 2006). What is meant by that can be

better described with the following example (as in Foxall et al., 2006). When a consumer

takes a trip to an exotic island this ‘situation’ can be accompanied by high levels of

utilitarian and informational reinforcement and at the same time high levels of aversive

consequences (increased cost, increased negative gossip etc.). However all of these are


287

expected to determine consumer behaviour at differing levels. This principle was indeed

verified by the present data. A possible connection between higher confusion indicating

lower levels of dominance (in the comparison between the two situations) was only partly

verified. High-technology market indicated lower levels of dominance however the

difference was very small and non-significant, which indicates that there are other

elements to determine the levels of dominance of a situation.

The next model examined consumer confusion as an intentional construct (and in that

sense identified the importance of personal characteristics and the individual apprehension

of situations). This conception of the BPM-I placed confusion in the position of the

consumer situation and examined the implications of this model. Different levels of

utilitarian reinforcement and perception of the behaviour setting scope were measured

however same levels of informational reinforcement were reported by different groups of

consumers. This is an interesting finding as it indicates that arousal (in the form of the

Mehrabian and Russell (1974) measurement) measures symbolic feedback on

performance which is mainly status related and possibly not knowledge/ understanding

connected. Other possible explanations for this relationship will be discussed in the

following sections.

A possible interaction effect between the market and confusion which are based on

theoretical arguments on the role of rule-governed behaviour and the idea that consumers

develop tracks in every-day situations has been examined. Evidence of this relationship

has also been found in the qualitative-pilot research in the form of consumer discourse:

Participant 2 (exploratory pilot research): No, no I don’t get confused when in a super

market. I have been shopping too long now to let it confuse me. No, I honestly quite enjoy

shopping.


288

Such effects were identified only for complexity confusion but not for similarity of

products (brands) which seems to be a more pervasive problem in retail environments.

Overall, this study has many findings and contributions which are described and discussed

in more detail in the following sections.

10.3. Theoretical Findings and Implications

The main findings of this study are summarised in this section of the thesis. In accordance

to the research objectives, apart from the arguments on the nature and measurement of

confusion, the section will extend to findings on its contextual treatment and effects and

the main theoretical models the BPM-E and BPM-I.

10.3.1. The Nature of Confusion

This study has extended the concept of confusion and placed it within the boundaries of

rule-governed behaviour (Skinner, 1969; Catania, 1986; Catania et al., 1990; Foxall;

1997a; Törneke et al., 2008; Foxall, 2013). In accordance with the categorisation proposed

by Zettle and Hayes (1982) the case of self-tracking can be used to describe the case of

confusion as a self- based rule, as this treatment implies the special relationship with the

state of affairs and environmental issues (Zettle and Hayes, 1982; Glenn, 1987). In order

to elucidate the actual nature of confusion, anomy, a concept introduced by sociologists

(mainly Durkheim and Merton) but which has been extended to a more general

psychological concept that indicates that ‘the world and the individual are adrift’ has been

used to place confusion in its actual position as a rule characterising the lack of other

rules. This position is in reality the case when clear rules are lacking and as described by

this research, market rules are either too similar or too complex.

Beyond the treatment of confusion at the extensional level as an overall situational

description in terms of brute facts, the findings indicate that the delimitation of


289

behavioural explanation and the personal level of explanation have been achieved by the

treatment of the construct at the less scientific level of intentionality. It is evident then that

intentional terms can be used to describe the consumer learning history and situation.

Moving from the ‘behavioural’ to a different theoretical background, namely the one

dealing with the nature of cognitions, emotions and emotional experiences (see chapters 2

and 5), the findings of this study might be of immense interest. Although, such

considerations are not part of the main theoretical framework of this research per se it is

worthwhile to explore the possible theoretical implications of this study for this area. The

effect of confusion on behavioural variables would have been treated in most research

studies (see for example Schweizer, 2004) as being mediated by emotional elements,

especially pleasure or dominance. The fact that confusion seems to have a significant

main effect along with these emotional elements could possibly be perceived as

corroborating the findings by Rozin & Cohen (2003a). In this case and following Rozin &

Cohen (2003a) the possibility that confusion can be perceived as having the characteristics

of emotions will be examined.

When considering the parameters that are usually measured to establish the nature of

entities as emotions these have been described as been (see also chapter 2): (1) the

language of emotions; (2) reflexive physiological activity (somatic and autonomic

reactions such as characteristic facial expressions (Ekman, 2001; 2003); and (3)

behavioural (e.g., approach and avoidance, ‘freezing’, and performance deficits or

enhancements). Rozin & Cohen (2003a) in their study very much based their findings on

the first two categories of measurement, language (students that acted as participants in

their study were asked to collect indications of emotions; in this case confusion was

repeatedly reported as belonging to this category) and facial expressions (confusion was

reported as an emotion in both the case of symmetrical and asymmetrical facial


290

expressions). The findings of this study could then suggest that the present analysis can

provide evidence to support the latter of the three criteria, that of determining behavioural

consequences in the form of approach-avoidance behaviour.

In this case the question remains: ‘would it be possible to argue that confusion is an

emotion?’ A clear answer should be prevented to be given based on the findings of this

study. At the most obvious level this study was not designed to pose an argument towards

that direction and additionally, an answer to such a question requires repeated empirical

investigations. An additional point would be that pleasure, arousal and dominance have

been described as the affective qualities of situations and environments (Russell & Pratt,

1980) and consequently these dimensions are either way not treated as emotions per se but

rather as dimensions with emotional inferences.

What can be possibly argued from the findings of this study is then that seen from this

perspective, confusion might have clear emotional implications in consumer settings,

which can be easily translated into behavioural consequences. If pleasure and arousal are

treated as being the ‘affective qualities attributed to an environment’ (Russell & Pratt,

1980) then confusion is likely to have the same emotional qualities (see also the

arguments that confusion can be described as a ‘cognitive feeling’ and that it admittedly

possess both informational and affective value in chapter 2) without however implying

that it can be placed in the category of pure emotions.

10.3.2. The Conceptualisation of Confusion

In order to achieve the measurement of confusion as a self-based rule, which describes the

lack of environmental/market rules in either a way that such rules are weak, unclear or

complex, an existing conceptualisation was used based on the idea that there are three

kinds of confusion, namely similarity (of products), overload (big variety of products and


291

information) and ambiguity (unclear information). The exploratory pilot test of this study

has managed to establish the ecological validity of these three kinds of confusion in the

retail settings of this study and in addition has established the validity of this measurement

in measuring actual traits connected to confusion rather than entities that can be more

meaningfully characterised as frustrating or annoying. The scale used has been validated

in a de-contextualised way in previous research (Walsh & Mitchell, 2005b; Walsh et al.,

2007; Walsh & Mitchell, 2010). This study acknowledges a problematic area with the use

of the characterisation ‘overload’ for what has been described as being faced with too

many products and information. The problem with this treatment is that a person can be

‘overloaded’ in the sense of facing ‘too much of a burden’ due to many reasons,

ambiguity of information can be easily perceived as one of them. However in order to

correspond with previous research (e.g. Sproles & Kendall, 1986) this naming has been

kept intact.

In this study, overload and ambiguity confusion were found to load together and form one

common factor named complexity confusion. This issue might be understood as a

problem with the dimensionality of the scale. The scale has however been tested in

previous research and it indicated elements of good dimensionality; it is suggested then

that this is a contextual issue rather than anything else. There are also firm theoretical

explanations for the common loading of overload with ambiguity confusion that will be

described. Specifically, conceptualised as one of the factors that characterised one of the

dimensions of information rate (Mehrabian & Russell, 1974), ‘complexity’ has been

defined as the ‘number and changes of environmental stimuli’. This conception clearly

coincides with the understanding of the factor that resulted from this factor analysis in this

study and has been named complexity.


292

The question then remains, how are the two constructs (similarity and complexity

confusion) resulting from this study dissimilar and what underlies their similarity? Can

these results theoretically be justified and connected with the understanding of confusion

as a self based-rule in this study? To start with both of these constructs can be considered

a case of anomy, rules developed to accommodate the lack of rules (McClosky & Schaar,

1965) as described in this research. High similarity of products is a kind of confusion that

creates homogeneity in a marketplace. It impedes behavioural responses by removing the

expected norms/rules (the expected differences in packaging/brand names etc.) that could

guide behaviour and thus acts as an aversive consequence of shopping. This impediment

of behaviour is based on everything being/looking similar. At the other end both overload

and ambiguity (complexity) impede behaviour due to ultimately different reasons which

are an increase in the complexity of an environment and actually have the opposite effect

(of making the environment more complex and disarrayed rather than homogenous). The

high correlation between the two constructs (oblique rotation was used in order to discern

between similarity and complexity confusion) indicate at the theoretical level that these

constructs indeed underlie the same idea, that of consumer confusion and anomy

(although the cause roots of behavioural impediment should be perceived as being

different).

At the practical level of managing consumer markets it indicates that managers should

remain very much focused on removing any sources of consumer behavioural impediment

because these seem to interact to produce a combined ‘difficult environment’ and act on

consumers’ inability to behave (choose products, shop etc.). Finding the ideal number of

products and information to be provided, which at the same time allows for the

discrimination between different alternatives seems to be the ideal case.


293

It is evident that in light of the results of this study, more (qualitative) research is required

in order to ensure scientific integrity and specifically delineate whether overload and

ambiguity confusion should be indeed treated as the same construct in future research, a

construct that can meaningfully point to complexity confusion. Taking as evidence this

study and the aforementioned theoretical arguments, all directions indicate that two

factors rather than three (the first indicating homogeneity and the other complexity) could

be a more meaningful way to measure and conceptualise contextual confusion in future

research.

10.3.3. The Contextual Treatment of Confusion

An additional issue that is worthwhile of a section in this final discussion chapter is

whether the treatment of confusion at a contextual, attitudinal level is, on balance, of any

research interest. Previous research has treated confusion at multiple levels with one of

them being the level of a personality trait- a personal proneness (Walsh & Mitchell, 2010).

This approach is de-contextualised and treats the market at an overall level. When

confusion was treated at a contextual level the measurement included just one market

situation for example the market of telecommunications (Turnbull et al. 2000). Other

contextual treatments are once again concentrating on one market and have also been

store and time specific approaches (Schweizer, 2004). The approach of this study can then

add an interesting statistic (refer to table 10.1) to the relevant discourse. This statistic is

the level of correlation between the kinds of confusion in the two different situations.

Table 10.1 Correlation coefficients between kinds of confusion in the two

markets

Similarity


Complexity


Similarity (Grocery) .323**

Complexity (Grocery) .394**


Source: this study


294

This interesting statistic was presented without extensive explanation in the analysis

chapter. As this is one of the studies to apply confusion in a context specific manner both

at the level of overall response to situations and as a propositional attitude towards a

situation, this correlation should be a good suggestion for future endeavours to indicate

whether confusion can add information when treated at the contextual level of every

market. A rather high correlation might indicate that confusion is better treated at the

personality level rather than the contextual. In this manner future research can have some

further guidance on whether this treatment is a justified and worthwhile approach.

The problem here is that to the best of this study’s knowledge there are no previously

published materials or relevant articles that could be used to lead the decision on whether

such correlation coefficients can be an adequate statistic to guide a decision of this nature.

It is evident that the relevant correlation coefficients indicate a medium positive

relationship between confusion levels in the two markets (both coefficients are below the

.50 value); it is however questionable whether such level of association would indicate

that confusion should be treated at the contextual level rather than as a personality trait.

Based on this fact, this study will adopt and extend on the approach taken by Mehrabian

and Russell (1974). In their 1974 book Mehrabian and Russell conceptualised the PAD

variables as potentially having both the properties of a personality disposition to react

with almost similar PAD levels in every situation (in that sense measuring emotional

reactions as personality traits) and a second one that treated the PAD as responses to

specific situations. The difference lies with the way the question is asked and the manner

the measurement takes place. A similar approach was presented in chapter 2, where

literature indicates that all states can have dual natures, one characterised as personality

traits and another one as contextual responses (e.g. Scherer, 2005).


295

To enrich this argument, specifically in this research, a correlation of .316** of the

reported levels of dominance in the grocery market with the reported levels of dominance

in the pc/laptop market was found. Similar correlation levels (e.g. .332** for approach-

avoidance behaviour) applied to the rest of the variables. In addition, in light of the

differing effect of complexity confusion on some of the PAD elements and approach

behaviour (interaction effects) between the two markets, there are obvious differences on

the ways confusion affects the two markets. In that sense the measurement and treatment

of confusion at the contextual level seems justified and is recommended for use in future

research.

10.3.4. The Extensional Model (BPM-E)

In the extensional behavioural perspective model, the discriminative stimuli set the

occasion for consequences, which can be classified into reinforcing and punishing

consequences of being in a setting. Figure 10.1 summarises this conception of the BPM

were utilitarian, informational reinforcement and the aversive consequence of confusion is

positioned at the right of the model and are described as the consequences of situational

exposure.

Figure 10.1 BPM-E

Source: Adapted from Foxall, 2013, p. 110.

Consumer situation

Consumer behaviour

Consequences

ReinforcementPunishment


UtilRInfR

Aversive consequenceof being in a setting

Confusion


296

This kind of reasoning, which used the extensional language of verbal behaviour, allowed

for the theorising of constructs at the overall level of responding to market conditions and

thus facilitated the attainment of the aim of the comparison of the two retail settings used

in this study, which seem to differ significantly in accordance to the principles of the BPM

(higher levels of reinforcement signal higher aminusa behaviour). In an attempt to

position the two research (retail) situations in accordance to the classification proposed by

the BPM contingency matrix the central contribution of the previous chapter has been the

compilation of a table with all means and range of the contingency situations as found in

previous studies and has conceptually tried to identify the relevant position of this study’s

situations.

The two levels of situational taxonomy that are proposed by the BPM (operant class and

contingency category) are described below and the relationship with the current situations

is discussed:

Level 1: Operant Class. The BPM model recognises 4 operant classes, accomplishment,

hedonism, accumulation and maintenance (Foxall, 1996). The main criteria are the

relevant levels of utilitarian and informational reinforcement. As data and means indicated

(see chapter 9) in terms of pleasure and arousal the grocery market can easily be placed

within the maintenance operant class. At the other end, the high technology market seems

to be more easily positioned beyond the maintenance class and shares the levels of

utilitarian and informational means with hedonism.

Level 2: Contingency Category. The concept of contingency category extends the concept

of the four operant classes and introduces the Behaviour Setting Scope as a way to

distinguish between open and closed situations. Openness is connected to being able to

define the setting, while closeness is identified in settings that are determined by agents


297

outside the individual. The levels of dominance reported for the two situations in this

study are relatively medium (5.30 for Grocery and 5.17 for the high technology market).

This is fairly a surprise as levels of dominance for such open situations like retail

shopping were expected to be higher (see also the relevant range of such situations in

previous studies of the BPM). The reason for that is that such shopping situations, which

are taking place in affluent, consumer orientated economies, marked by high levels of

discretionary income, open settings are supposed to be commonplace and consumer

choice is supposed to be sustained by competition among providers (Foxall, 1992b). It is

however relevant to mention that levels of dominance for the routine purchasing have

been found to range between 5.5-6.4 and have a mean of 5.9. These means are relatively

higher from the ones found in this study.

Regarding the comparison between the two specific situations, the grocery market

indicates a slightly higher mean than the high technology market but this difference is

neither big nor significant. It is interesting to state that the levels of reported confusion

would justify (significant) lower levels of dominance for the high technology market,

however it seems that because there are other reasons that influence the closeness or

openness of a situation there seems to be no further implications. It has been proposed that

levels of closeness-openness are influenced by the number of accessible alternatives from

being to the setting and it seems that both of these consumer oriented situations offer

alternatives.

Further to the above established measures of utilitarian, informational and behaviour

setting scope, this study introduced the idea that the overall levels of similarity and

complexity confusion are part of the aversive consequences of shopping and has measured

them in both markets in order to identify their extend. Similarity confusion has been found

to have similar levels in the two markets while complexity confusion is higher in the high


298

technology market. According to the principles of behaviourism any levels of one

dimension (meaning reinforcement- aversive consequences) can be accompanied by any

levels of the other as there are situations (buying an expensive car) characterised by high

utilitarian reinforcement (good quality) and high aversive consequences (increased cost),

however the levels of behavioural responses expected are determined by the

reinforcement received.

10.3.5. The Intentional Model (BPM-I)

The Behavioural Perspective Model has been extended in this study to include intentional

constructs in the form of an individual propositional attitude and more specifically

confusion in the model. In this case confusion is described in the form of the individual

beliefs (avoiding a more cognitive language) and has been used to indicate that individual

rule-making can modify situational contingencies by increasing or decreasing the

reinforcement consumers receive and the way consumers perceive the behaviour setting

scope at the personal level of explanation this time. In order to adopt this approach a less

‘scientific’ route is adopted (Foxall, 2013) where beliefs and intentional constructs are

positioned at the left of the model as in figure 10.2, mainly indicating consumer learning

history and acting as the consumer situation that signals appropriate responses.

Figure 10.2 BPM-I

Source: Adapted for this study from Foxall, 2013, p. 116.

Consumer

situation

Consumer


Reinforcement

Punishment

Intentionality

Belief/ Propositional Attitude = Individual Confusion

UtilR

InfRBehavioural Setting Scope


299

Based on this understanding the following relationships have been explored:

a. Confusion- Utilitarian Reinforcement/ Behaviour Setting Scope/

Behavioural Responses

Levels of both similarity and complexity confusion have been found to have a negative

relationship with the utilitarian reinforcement that groups of consumers perceive as

receiving from situations. This is also relevant for the behaviour setting scope with

confused consumers perceiving that their access to reinforcers (mainly the choice of

products in this case) is being regulated by other agents. The interaction effect of the two

markets when it comes to complexity confusion should be reminded however, this is the

case where pleasure, dominance and approach behaviour have been found to be more

highly negatively influenced in the high technology market (market characterised by

overall lower levels of experience).

On these grounds approach and avoidance behaviour have been found to correspond

accordingly to the lack of perceived reinforcers. It should be noted that especially

avoidance behaviour seems to be more highly influenced by both kinds of lack of rules

and in both market situations.

b. Confusion- Arousal/ Informational Reinforcement

The relationship between confusion and arousal has been described as a rather unexplored

matter that is worthy of deeper investigation. Elements like the proposed relationship of

information rate with arousal (Mehrabian & Russell, 1974), the additional findings from

psychological research that found evidence that confusion is to be understood as an

unaroused state (Russell & Mehrabian, 1997) and further ambiguous results from

subsequent research of consumer behaviour (Donovan & Rossiter, 1982) complicate the

matters. The findings of this study confirm that overall confusion should be indeed

perceived as an un-aroused state. In reality, the conceptual development and hypotheses of


300

this study propose the possible reason for this lack of relationship. In case arousal is

perceived as measuring the feedback on performance which has until now connected to

levels of self-esteem/ status associated feedback, this is the kind of status that comes from

symbolic reinforcement (e.g. Foxall & Soriano, 2005). It is possible then that either

confusion does not create such kind of symbolic/ status related punishment or that simply

arousal is not in a position to measure the kind of feedback on performance produced by

confusion which is possibly more of a ‘personal’ feedback on the levels of understanding.

Regarding the relationship between confusion and arousal, this study will also bring to the

fore a proposal based on an alternative explanation. This explanation is still based on the

conception of arousal as a measure of feedback on performance, however elements of

attribution theory (mainly the distinction between internal and external attribution) that

can be used to explain the relationship between confusion and arousal.

Kelley & Michela (1980, p. 458) argue that there are many kinds of ‘attribution theory’ as

the term overall refers to the perception of influence of a cause on a behaviour. The

common theme underlying these theories is the way that ordinary people explain events

and make sense of the world the way they do. The theme of internal and external

attribution (or locus of control) is recurring in attribution theory and describes the

attribution of an event to either internal (within the individual) factors, or contrary to

external, possibly environmental factors (Kelley & Michela, 1980, p. 487). It is then

suggested that consumers as naive psychologists (Heider, 1958) have the tendency to

attribute confusion either internally (to the incompetence of oneself to understand and

interpret environmental stimuli efficiently) or externally (to the incompetence of the

industry to provide a clear and unambiguous environment for them) - see also Mitchell et

al., 2005 on the topic of confusion and attribution. Specifically, it is suggested that

attribution can shape the relationship of confusion with arousal in the following manner:


301

In case consumers perceive confusion as their personal incompetence this will have a

negative effect on the perceived levels of personal performance, thus lower levels of

arousal (informational reinforcement) are expected for confused consumers than non-

confused consumers (providing that consumers attribute confusion internally and thus it

does have an effect on the informational reinforcement they receive).

In case consumers attribute confusion externally no effect of confusion on arousal will be

identified, thus confused consumers are expected to indicate the same levels of arousal

with non confused consumers (providing that confusion will be attributed externally and

thus it has no effect on the levels of informational reinforcement). This theoretical

argument and the findings of this research might dictate that confusion is not perceived as

providing any feedback on individual performance thus it might be an indication that

confusion is perceived as more of an external characteristic of a market and thus this

situation is attributed ‘externally’ by consumers. Theoretically, this finding is interesting

because it extends the possible theoretical connection of attribution theory with confusion.

An initial proposal on this matter has been that attribution theory serves to determine the

consequences of confusion (Mitchell et al., 2005). According to this rational, the more

consumers attribute their confusion to external sources, the greater the suggested effect on

company-related consequences like negative word of mouth for example. This study

proposes then a possible extension of the relevance of attribution theory to the effect of

feedback on performance which is worthwhile of further investigation.

Further to that, the overall relationship between arousal and similarity was negative, small

and not significant. The small, non-significant but positive relationship of arousal with

complexity confusion might indicate the possible theoretical connection of the complexity

confusion construct with the information rate of an environment. Previous research has

found mixed results on the relationship of information rate with arousal. As one of the


302

cautions of this study it is then suggested that future research could examine the

relationship between information rate and confusion and place it on better grounds.

Further theoretical arguments are then necessary to understand the concept and

measurement of confusion. As such, further measurements of the construct could abandon

the well-grounded treatment of consumer confusion based on its antecedents (as

summarised in the conclusion of chapter 3) and could focus on alternative characteristics.

Characteristics as the ones proposed by psychological research could offer a breakthrough

to this issue. For example elements such as the perceived intense added effort or attention

required or the sense of goal obstruction, all summarised by previous studies (e.g.

Ellsworth, 2003) to characterise the ‘qualia’ of confusion—the individual instance of

subjective, conscious experience—could find a role in future research in an attempt to

develop a measurement for the construct based on an alternative perspective. This

perspective could have less of a relationship with information rate and could add value on

the debate on the relationship between confusion and arousal.


303

The BPM and the MR Approach

As discussed in both chapters 8 and 9 (analysis and further analysis), contrary to some

evidence from previous research (Donovan et al., 1994) which have been critical on the

way the measurement and specific conceptualisation of arousal and dominance represents

emotions felt within retail environments, this study confirms that overall the original

scales indicate good reliability and validity in retail settings. One item of dominance

(awed-important) and one of arousal (sleepy- wide-awake) had to be removed from the

corresponding measurements. Approach-Avoidance measurements also signified good

reliability and uni-dimensionality (although it is relevant to remind that especially for one

of the items of approach which concerned the time spent in a retail situation some

modifications had to be applied, in order to reflect consumers’ understanding of the

specific settings).

Confusion as a variable has a rather small but significant contribution (ranging from

0.5%- 2.5% unique R square change) to the overall models when the three behavioural

variables acted as the depended variables. This small but significant contribution has been

judged as sufficient for the requirements of this study. The greater contribution is with

avoidance and aminusa and the contribution is much less with approach. It should not be

forgotten that confusion was measured at the level of responding to stimuli and has been

treated as an overall belief about specific markets; thus this small contribution should not

be perceived as insignificant. It indicates that elements of the overall understanding that

consumers develop for specific markets do have an effect on behavioural variables and

work along with the reinforcement received from these situations to have an effect on

behaviour.


304

The main aim of this study has not been to identify variables that could increase the

explanatory power of the MR model but rather variables that as aversive consequences

and in accordance with the BPM, work along with utilitarian (pleasure) and informational

(arousal) reinforcement and the behaviour setting scope (dominance) to determine

consumer behaviour. This study has been successful at that aim. It has above all strived to

identify a way to, theoretically and based on appropriate research practices accommodate

and examine situations that the contextual and the intentional stance can hold together and

explain behaviour. Based on Foxall (2013) this can be achieved due to the property of

rule-governed behaviour to be described in an extensional and intentional language and

confusion has been defined in this study as an ideal case for such an endeavour.

Overall, levels of utilitarian reinforcement (in the form of pleasure) seem to be a very

important element that determines behaviour in choice settings. In accordance with

previous research and although in this study the original Mehrabian & Russell (1974)

scale was used, dominance has been found to have no contribution to the overall model

when applied to consumer settings/markets. The implications of this finding will be

discussed in more detail in the section that follows.

10.3.6. The Role of Dominance

Chapter 5 of this thesis on the Mehrabian and Russell (1974) model included a theoretical

section dedicated to one of the elements of the PAD, dominance. The reason behind this

‘special’ and extended treatment of dominance has been its ambiguous nature and

inclusion in the model. Arguments regarding its removal from the PAD model derive both

from psychological research (Russell, 1980) and its possible unsuitability as one of the

variables used to characterise emotional reactions (based on the propositions put forward

by the dimensional theories of emotions) but also its possible inadequacy to explain

behavioural responses in retail settings (Donovan & Rossiter, 1982). As a


305

counterargument Soriano & Foxall (2006) propose that dominance has been found to

discriminate well between open and closed situations as defined by the Behavioural

Perspective Model. In addition, in cases a structured approach based on the situations of

the BPM contingency matrix is adopted (Foxall & Greenley, 1998) all three elements of

the PAD seem to indicate more consistent results when considering their relationships

with the behavioural variables (Foxall & Soriano, 2005).

Regarding the levels of dominance between the two shopping situations examined, these

situations indicated same levels of dominance and this has significant implications both

for theory and for the management of retail settings. In addition, based on the alternative

model BPM-I confused respondents have been found to report lower levels of dominance,

and in that sense at the personal level, settings are perceived as more closed.

Regarding the lack of dominance capacity to determine behavioural responses, it seems

that the effect of the situational consequences (reinforcing and aversive) is stronger in the

case of retail settings. The following four observations have been developed and will be

discussed in terms of this aspect and the fact that dominance was found to be deficient in

defining Approach-Avoidance (Aminusa) behaviour along with pleasure, arousal and

confusion:

1. Foxall & Greenley (1998) attributed the insignificant results of dominance in

previous retail research to the allocation and examination of arbitrary situations. This

has been the result of the lack of a systematic approach to the way situations are

structured which has turned the prediction of differences among them in terms of

behavioural responses difficult. Dominance has subsequently shown good signs

regarding its capacity to determine behaviour when all eight diverse situations

proposed by the BPM are measured. It should be taken into consideration that the


306

present study has due to the very nature of the measurement and situations necessary

(as explained in the conceptual framework) fallen into this ‘fatal’ unsystematic error

(according to Foxall & Greenley, 1998). Thus, it is possibly the lack of the entire array

of situations which does not allow for all different levels of dominance to be present

that might cause dominance not to work in cases that not all of the eight situations are

examined.

As much as this explanation seems to be plausible, the problem remains that not all

everyday consumer reality and equally not all research situations to be examined are to

include all eight contingency categories as proposed by the BPM. The ‘systematic

approach’ of the BPM although theoretically useful it cannot find exact application in

every research. In most research cases it is desirable that less than eight situations or

most commonly only one situation is to be studied. The definition of a theory will be

brought to the fore to explain the reasons that such a treatment seems to be

problematic. Blaikie (2000, p. 142-143) compiled several different definitions of theory

and have resulted to the following composite meaning. A theory is then:

‘a related set of statements about relationships between concepts with a certain level of

generality which are empirically testable and which when tested, have a certain level

of validity’ (Blaikie, 2000, p. 142-143).

If then dominance cannot determine behaviour when only one or few situations are

tested this is a challenging area because it questions the generality of this theory (as has

already been done in the literature of consumer behaviour) or that a more accurate

classification of consumer cases that this construct should be included or not should be

constructed.

2. The second point is very much related to the above reasoning but extends the

argument based on research related with the dimensions of emotions. Morgan and


307

Heise (1988) administered emotional words to college students and asked them on their

ratings of relevant dimensions. One of these dimensions was dominance which was

conceptualised as potency in their study. The relevant measurement included only one

bipolar phrase: big, powerful versus little, powerless. Their results indicate that

potency indeed differentiates between the negative emotions of fear (terror) and anger

(fury) however little evidence was manifested for the effect on any other negative

emotions or especially on any positive ones because all positive emotions generate a

sense of powerfulness. It is then concordant to say that because most retail

environments are not the kind that would produce intense unpleasant emotions (e.g.

Donovan and Rossiter, 1982 could not find enough negative shopping occasions in

their research), future research should focus on creating a classification of consumer

situations that dominance could be relevant or not. It seems then that one possible

explanation for the inability of dominance to explain approach-avoidance behaviour is

that retail settings are so designed that on average produce a pleasant consumer

situation, which do not allow for dominance to really differentiate or have an impact

upon.

3. Another plausible explanation for this research aspect should evidently extend to

the definition of consumer situation according to the BPM-I. The consumer situation

has been deconstructed in this case and includes intentional elements other than the

behaviour setting scope. In this explanation dominance is then not a defining aspect of

the model but has been conceptualised as a possible consequence of being in different

settings. Are we, the theorists, then allowed to argue that: 1) possibly dominance is not

a necessary or better a ’defining’ characteristic of the BPM-I and even that 2) possibly

the BPM-I can act as a better conceptual edifice of some consumer situations? The

answer to such matters cannot be clear yet. The answer might be ‘yes these argument


308

are valid’ or ‘no this is not the case’ in other instances. In order to have a definite

response more intentional models need to be examined and more intentional constructs

need to be added to the model.

4. Finally this lack of capacity to determine behaviour should turn our focus once

again to the literature of emotions and cognitions and the explanation proposed by

Russell (1980) (see chapters 2 and 5). In that sense a possible mediating effect of the

emotional dimension of pleasure (and possibly arousal) on the relationship between

dominance and A_A might be assumed. According to Russell (1980) dominance was

excluded from the bipolar (pleasure and activation) affective environmental qualities

on the grounds that it requires cognitive intervention. This requirement for cognitive

intervention has been seen as its defining difference from the emotional dimensions of

pleasure and arousal. Keeping this fact on the one hand and the original definition and

process of the process to identify mediation by Baron and Kenny (1986) at the other,

then the conclusion of mediation is effortlessly reached.

In Baron & Kenny’s (1986) description a mediator is defined as a variable that explains

the relation between a predictor and an outcome variable. The criteria to establish

mediation then dictate that: a relationship should exist between the predictor and the

outcome variable, the predictor and the mediator variable and the mediator and the

predictor variable but the mediator variable has to reduce (partial mediation) or

eliminate (total mediation) the link between the predictor and the outcome variable

when entered into the relationship. Bringing that in terms of this specific research a

relationship between dominance and approach/avoidance exists and dominance has

been found to have an effect on both pleasure and arousal in past research- Ward &

Barnes, 2001- a correlation between the variables exists in this research- however this

relationship is eliminated when the variables pleasure and arousal are entered into the


309

regression. Future research should examine the possibility that there is a mediating

effect of pleasure or/and arousal to the relationship between dominance and approach-

avoidance in consumer environments. In order to examine such a relationship a

different theoretical perspective is necessary from the one in this study and would

further require the application of relevant mediating statistics. The significance of the

present study lies in the way in which the study’s findings and research questions are

illuminated by using behaviourism, the BPM and the philosophy of intentionality as a

major frame to enlighten the phenomena and thus this topic is beyond the interest of

this study.

However, a query caused when adopting this theoretical approach is the examination of

the reason that confusion seems to work along with pleasure and arousal to determine

behaviour while dominance is not. When following the aforementioned logic and

previous theoretical arguments confusion can similarly to dominance well be a

construct requiring cognitive intervention. The only plausible answer to that is that

specifically in consumer environments confusion might have more intense emotional

implications than dominance. Our time is the era of marketing and huge importance

has been placed to consumer rights and well-being. Thus confusion which is an

aversive consequence of shopping might have stronger emotional implications than

dominance in such contexts.

Following this examination of the issue, the exclusion of dominance from future

endeavours is not suggested. Rather, seen from multiple theoretical perspectives

dominance has a significant role to play especially in emotional ascription but also in

behavioural understanding. It is however advised that more studies should specifically

focus on the relationships and effect of this variable in one or more environments. It is

strongly proposed that further situational classifications could delineate the role of


310

dominance especially when distinguishing between pleasant and unpleasant situations

(similar to the differentiation of the effect of dominance on positive and negative

emotions).

10.4. Managerial Implications

The need to address situational influences when creating retail settings/stores has been a

conventional practise for retail managers for years now. By adopting the BPM and the

specific approach of this study, more specific advice on situational effects like the levels

of pleasure, arousal, dominance and confusion that seem to encompass individuals’

emotional responses to social and physical environments (see also Foxall & Soriano,

2005), are offered to retailers as an advice on how to form retail settings.

Although such advice is not new to retailers and the PAD approach has been used before

as a meaningful way to study retail settings, the approach of the BPM contributes to an

alternative manner that retailers could perceive their environments. The conceptualisation

of arousal for example as the level of informational reinforcement provide retailers with

evidence that along with elements that increase utilitarian reinforcement (good product

and service quality), other elements of the markets that increase social prestige,

performance feedback and self-esteem (like for example either the nature and use of the

product- a PC brand carries with it more informational implications than groceries or the

implementation of a loyalty/ point collection scheme) are important determinants of

positive consumer behaviour but should not be so strong as to detract from the pleasure

responses generated by such environments.

Turning the focus on the central variable of this study, consumer confusion, it has been

indicated that by reducing the levels of complexity and similarity confusion and

conversely increasing clarity, pleasure and dominance could be a major source of

competitive advantage in any market, but particularly in those markets where confusion


311

has been proved to be an aversive consequence like the grocery and high technology

markets in this study.

Considering the limited marketing resources, organisations need to consider how best to

deploy them in order to assess and reduce confusion causing activities. In order to

implement this task, the conceptualisation of confusion in this study provides marketers

with guidance on the areas where attention may be required. In this manner, this study has

established that: 1) confusion can indeed be described as a situation characterised by the

‘lack of market rules’ that impedes consumer behaviour and 2) that it is essential for

marketers and retailers to check and control both of the two dimensions of confusion

(similarity and complexity) and further examine the ways that these different dimensions

influence consumer behaviour in retail settings. As a first, essential recommendation then

marketers ought to systematically identify sources of stimuli similarity and complexity

and try to rectify them.

Specifically, the dimension of similarity confusion concerns aspects of product and brand

similarity and has been established as an issue in markets with varied levels of experience

and other characteristics (no interaction effect between markets and similarity confusion

was found). This factor seems then to be influencing consumer vulnerability (in terms of

behavioural and emotional dimensions) in any market and it is independent of the levels

of market experience or other market characteristics. This finding should concern

especially policy makers and marketers of the grocery market who can use the

measurement of this study to examine further whether consumers actually perceive very

little differences between their own and competing brands. If this is the case, they should

as a first step seriously examine and reconsider their brand positioning and product

differentiation policies. As retailers wish to retain the clarity of their environments, lucid

positioning and differentiation of products can increase the selling power of products and


312

provide rules to facilitate consumers’ decisions. At the other end, through clear

positioning and differentiation, product managers and sales departments can increase sales

but also increase their company’s purchasing and negotiation power over retailers.

It is true that plenty of studies have been implemented in the past, concerning especially

the existence of me-too, ‘copycats’ or look-alike products and brands and these studies

have been especially focused in the grocery industry (Balabanis & Craven, 1997). The

finding that similarity confusion is an equal problem in both markets in this study might

be especially so due to the exact nature of grocery retail environments. Grocery stores are

the kinds of environments where all products are positioned very close together (consider

and compare a grocery store in comparison to an Apple store) and this fact might induce

the strong effect of the similarity confusion. Thus, although in a high technology market

consumers lack experience, this lack of familiarity might be compensated by the attempts

of the industry to provide more structured retail environments and better differentiated

products. Extending on this same argument marketers and store designers ought to 1) pay

more attention to the design of grocery stores, 2) to revise category management

techniques which focus on shelf arrangement and 3) to re-examine the practice of constant

introduction of new products which simultaneously make use of very similar marketing

strategies as their counterparts.

The second dimension of confusion, complexity, deals with issues like the vast variety of

products/ stores and offers in the market-place but the problem does not end there; rather,

unclear and ambiguous information is part of the complex environment created. Marketers

should try to reduce this complexity in multiple ways. To start with, they should rethink

the tactic of constant new product launches as a way to gain instant profit. Information

provided should be clearer and the different market sectors (with the help and support of

policy makers) could potentially agree on some common guidelines or official definitions


313

for the use of confusing or constantly changing terms on packaging. Policy makers could

also establish a free public body in the same logic as ‘Which’ organisation which can

provide guidance and support to consumers when faced with important choices.

The additional finding that complexity of markets is very much correlated with the

dimension of product similarity should concern retailers and marketers. Although the

causality of this strong relationship is difficult to be established this finding indicates that

the optimum market environment should be considerate of both complexity and

homogeneity which are both causes that impede consumer choices. This optimum market

environment is however difficult to be achieved and requires extensive effort and further

research from retailers.

In this study confusion has been established as a problem in retail settings and seems to be

especially connected to avoidance behaviour in all markets. As such retailers have an

additional benefit to strive for the creation of the above mentioned ‘optimum

environment’ because in such retail environments consumers would be more willing to

leave, or feel unfriendly. As already described, environments might become less

confusing if better differentiation strategies, clearer packaging, store sign-posting and well

educated staff are used in stores so that environments and products become clearer, less

ambiguous and easier to understand.

Beyond the issue of confusion, this study has established that a measurement scale like the

PAD can meaningfully identify differences in the contingencies of different markets. The

use of the PAD and behavioural measures in cases retailers wish to identify the levels of

reinforcement or the levels of openness/closeness and behavioural consequences of a

market situation has been established following this research. Retailers and marketers

could well use the PAD scale for comparison of the levels of utilitarian and informational

reinforcement which are induced by different alternatives of products before any new


314

product launches or adequately test retailers’ concept stores so that they can measure the

relevant responses induced to consumers by these situations. In that manner a more

informed decision-making on whether to launch specific products or on the way to create

a retail environment can be made based on the results of their research.

It is then evident that the findings of this study have implications for practitioners

(marketers, retailers and policy-makers), who will all benefit from following some of the

guidelines that this study offers.

10.5. Contributions of this Thesis

Based on its findings, this study has attained its research objectives and has many

theoretical and practical contributions. The main contributions of this study are twofold.

Most importantly it adds to the extant literature on the BPM. Second it offers a furthering

of the concept of confusion towards an alternative philosophy which enables many future

further explorations. More specifically, the contributions of this study in the respective

areas are presented below.

10.5.1. Contributions by Chapter

Chapter 2: Chapter 2 reviews theoretical and practical studies of general psychology and

more specifically discusses the ways that confusion has been dealt in that part of the

literature. This is to the best of this study’s knowledge the first attempt to summarise such

diverse sources and adopt such an angle and focus to the study of confusion.

Chapter 3: Chapter 3 has expanded on the concept of confusion and the ways it has been

explored in consumer behaviour. This chapter offers a comprehensive updated description

of the state in the literature. Starting from concepts like variety-seeking but also optimum

stimulation level, the way that these have led to the study of confusion has been

exemplified. Previous conceptualisations of confusion are described and most importantly


315

the demarcation of terms like confusion and other psychological states (cognitive

dissonance, perceived risk and frustration) are discussed. This chapter contributes to the

literature on confusion through the comprehensive approach taken.

Chapter 4: Chapter 4 is mainly a theoretical chapter which along with chapter 5 lays the

foundations for the conceptual framework. The two chief psychological frameworks of

conducting research, cognitive and behavioural psychology, are explained identifying

their theoretical strengths and weaknesses. The chapter deals with the theoretical basis of

the Behavioural Perspective Model (the contextual stance/ extensional language) and

moves one step further to support the importance of intentional terms in the study of

human behaviour as proposed by the work of Foxall (Foxall, 2004; 2007a; 2007b; 2008;

2013).

Chapter 5: The framework proposed by Mehrabian and Russell (1974) (discussed in this

study as the MR approach) has evolved into one of the most useful approaches to

examining environmental and situational influences on behaviour. Chapter 5 places the

model among other theories of emotional ascription, summarises the initial work on the

development of the model and extends to recent approaches utilised for its application.

The ways this framework has found application in the examination of the BPM is

discussed.

Chapter 6: The conceptual understanding developed in chapter 6 is one of the main

contributions of this study. The chapter provides a novel proposition on the nature and

treatment of confusion. It then places this idea in the frameworks of the extensional and

the intentional behavioural perspective model (BPM-E/ BPM-I). Consumer confusion has

been proposed to be a case of self-based rule, better understood as a rule about the lack of

rules (in this case weak (–similarity confusion) or complicated (–complexity confusion)


316

rules that can impede behaviour), and more specifically a case of track. Tracks are rules

resulting from the state of affairs (or environmental affairs). This construct is then placed

and explained in terms of both the aforementioned models, specific research hypotheses

were developed and research was conducted based on this novel framework.

Chapter 7: Chapter 7 on research methodology placed this research into the philosophical

boundaries of intentional behaviourism and methodologically described it as a quantitative

study. A quantitative survey possesses a prominent role over the qualitative element,

which has a supportive/ pilot role to facilitate the process of questionnaire development.

The results of the small scale exploratory pilot study were incorporated/ presented in this

chapter. These findings, enriched with appropriate theoretical arguments, indicate the

more suitable and free from other traits conceptualisation of confusion as used in this

study.

Chapter 8: Chapter 8 is concerned with the analysis of the survey/ quantitative data.

Several different uni- bi- and multi-variate statistical tests have been used in the

exploration of the data.

Chapter 9: This chapter offers further data insights and juxtapose present findings with

previous studies. The possibility of establishing socio-demographics as a viable

segmentation basis for confused consumers was examined. Further to that, the results of

this study were compared with previous findings. Tables 9.7 and 9.9, where all findings of

the studies of the BPM have been compiled, has been helpful in describing that the high

technology market seems to differ considerably from the contingency category 7 ‘routine

purchasing’ and especially in terms of informational and utilitarian reinforcement it is

characterised by the operant class of ‘hedonism’.


317

Chapter 10: This is the final chapter of this thesis. Findings are discussed and interpreted

based on the main theoretical assumptions of this study, the principles of the BPM.

Additionally, results are illuminated based on other theoretical approaches. In light of this

chapter several alternative venues for research are proposed.

10.5.2. Objective Driven (and Overall) Contribution

Following this examination of the contributions of each chapter of this study the overall

contribution should be understood as having multiple levels in accordance with this

study’s objectives.

1. This has been the first attempt to explain an interesting phenomenon like confusion

from a behavioural perspective and apply the principles of this stream of research

(especially rule-governed behaviour) to the understanding of its study.

2. Although the literature on consumer confusion lacks deeper knowledge of the

emotional implications of confusion (Mitchell et al., 2005), the PAD measurement of

emotional reactions to environments has been used in its alternative conceptualisation

of the utilitarian, informational reinforcement and behaviour setting scope of

situations/settings. This is a proven capacity of the measurements of the PAD to

represent these concepts as proved by research on the BPM contingency matrix

(Foxall, 1997b and all subsequent research). Based on these ideas relationships

between confusion and the PAD and behavioural elements have been described. The

theoretical positioning of this thesis allows for the elucidation of the relationships in

an alternative theoretical manner. For instance, the relationship between confusion

and dominance has taken the argument a step further by extending the idea that

consumers feel helpless when confused and adding the dimension of closeness

(meaning that other agents have more power over the self in a situation) in the

understanding of this relationship.


318

3. This study further provides theoretical support and validation of a recently developed

scale of consumer confusion17

through its application and testing in a contextual

manner and in a different array of consumers.

4. Corroborating research objectives, both similarity and complexity confusion along

with pleasure (utilitarian reinforcement) and arousal (informational reinforcement)

have been proved to determine especially avoidance and aminusa behaviour. This

indicates that the integration of confusion in both the BPM-E and BPM-I, as either an

aversive consequence or a summate of the consumer situation and learning history

indicating the personal level of explanation is supported by the data and acts as one of

the most important contributions of this study. The implications of this finding not

only for the study of the BPM but also the extant theory on consumer emotions have

been analysed before.

5. This study has then achieved the exploration of the two alternative models that are

offered to researchers to study consumer behaviour, the BPM-E and the BPM-I.

6. Methodologically, enough evidence is provided to demonstrate the responsiveness of

the original PAD and behavioural scales to retail environments and an online survey

methodology.

10.6. Limitations and Directions for Future Research

Wells (2001) argues that both marketers and government officials need theory that they

can trust and need to know the way that ‘real consumers make real decisions’. This study

has adopted all measures in order to satisfy this demand by producing reliable and valid

results from real consumers (see chapter 7 on sampling techniques and methodological

choices to ensure validity and reliability, and chapter 9 on comparison of the present

17 Confusion scale developed by Walsh et al., 2007.


319

results with previous studies). As one of the few attempts for the incorporation of

intentionality (and especially individual intentionality) into the Behavioural Perspective

Model many opportunities for future research are offered (both in terms of replications

and theoretical and empirical debates and extensions).

It might be objected that one of the limitations of this study lies with the fact that

consumers were asked to report their emotional and behavioural responses as reactions to

descriptions of situations and not as responses to actual situations. This should not be

perceived as a limitation of the study but rather an informed choice based on some of the

theoretical implications of the model and the understanding developed. This is once again

the same case as described before18

against the use of more technologically advanced

approaches like slides/photos/videos, because these impose a specific setting to

consumers. The use of actual/real situations could force consumers to respond to the

specific circumstances they face rather than express their overall understanding in terms

of a learning history. That means that if a participant was asked to answer the

questionnaire outside a grocery store and having already spent for example twenty

minutes to ensure a parking place, then this participant would be more likely to report

more negatively for the case of grocery shopping rather than when answering

uninterrupted at any other time of the day- where a learning history and intentionality

would be put to the fore. As Donovan et al., (1994, p. 292) explain the problem with in

store emotional measurements is that such measurements might reflect feelings brought to

the environment rather than induced by that. Thus the use of shopping descriptions is

essential and desirable for the requirements of this study. In case future research intents to

re-focus on the connections between the MR model and confusion, a re-examination of

18 Refer to chapter 7.


320

the variables in actual situations or following real shopping experiences would be useful

in determining the way the model responds in such a different theoretical and research

background.

10.6.1. Limitations

Notwithstanding the above noteworthy elements that establish the ‘goodness’ of this

research, the approach can be said to be somewhat limited in the following respects:

1. Use of the descriptions of two shopping situations (in the future more situations

could increase the results’ generalisability). The fact that only two shopping situations

were used has been a problem at two levels: a) this research is based on two situations

which might point to some limited scope. Some of the results might require further

elaboration and testing in more situations. For example the role of dominance in

behavioural determination as discussed above might differ when more situations are

examined. More markets could further provide a better depiction of confusion as an

extensional and intentional construct and could allow for a more meaningful

comparison of situations. b) Both of the situations examined are part of today’s open

market and although are expected and actually found to differ significantly, these are

still rather close to each other in terms of the levels of PAD, confusion and experience.

The expectations on the results variability cannot be the same as when eight diverse

situations as proposed by the original BPM are examined. For example situations like

being at work related conditions on the one hand and being at an exotic island on

holidays at the other are the kind of situations that are more likely to both create

stronger and more diverse emotional experiences (both positive or negative), which

would facilitate the process of establishing differences and bigger diversity in the

dataset. Bigger variability in the dataset could also result in higher and stronger

correlations among the constructs, which might be somehow restricted in this case.


321

The case that consumer/shopping situations on average are not the kind of contexts that

could create extreme emotional experiences has been discussed in the past and several

researchers have proposed some measures, like the application of appropriate scales

which have been developed specifically for consumer contexts- e.g. Richins, 1997;

Huang, 2001 to solve this issue rather than measures of general psychology (this

discussion should of course leave aside cases of consumers’ intense emotional

reactions like extreme anger or rage- Bougie et al., 2003; McColl-Kennedy, et al.,

2009).

2. Respondents: Overall the sample characteristics were based on a quota sample in

accordance with the UK population mainly in terms of age, gender and other socio-

demographics. The main aim has been to reach a cross-section of diverse consumers

with varied experiences in the market place which would establish more variability in

the answers provided. However, several minorities with specific characteristics like for

example less advantageous consumers or those who might not have an internet access

might not be part of this sample. Although as discussed in the methodology chapter the

overall internet reach especially in the UK is at a more than satisfactory level, the

problem of reaching specific populations remains, but the influence in this research

should be perceived the same as with any other form of collecting survey data. Finally,

the conduct of the research in one country can be considered as limiting the

generalisation of the findings across other populations, cultures and countries with

different retail and market conditions.

3. Measurement of actual behaviour: This study has used consumer verbal

responses to consumer situations. However more research would be desirable to

compare reported responses with the actual consumer behaviour. One of the most well

known research approaches that can facilitate report of actual behaviour is the use of


322

diaries (Bolger et al., 2003; Alaszewski, 2006). Although the behaviour in dairies is

still ‘reported’ and not observed, it is still reported much closer to the actual

experience. Participants of future studies could keep a diary of their emotional and

behavioural responses during or following repeated shopping occasions. This pattern

of the emotional values of these consumers would then be interesting to be correlated

with their reported confusion levels. This approach could help the creation of a more

accurate depiction of consumers’ emotional and behavioural responses, which would

be based on repeated patterns of experiences. Such a study is very difficult to be

implemented due to the level of resources and participation required which is difficult

to be achieved (a main problematic area with diary studies has been described as the

low response rate achieved- Alaszewski 2006).

10.6.2. Directions for Future Research

Many areas which arise from this study merit further research and these will be described

in more detail in this section. On the grounds of this study’s limitations and following the

overall evaluation of this research, measures of the actual consumer behaviour would be

an interesting avenue for future research as described in the previous section of research

limitations.

Further to that, in support of the replication stream of research in marketing, the present

study should be replicated with additional samples, retail/shopping situations and cultures

in order to determine if the findings are consistent in other settings. As countries and

cultures tend to differ a lot in terms of the arrangement of their retail settings, the maturity

of the retail system but also in terms of the value placed on consumerism, it would be

interesting to learn more about the role of consumer confusion in such different consumer

environments.


323

In terms of the implications of this research for the study of the BPM, the results indicate

the relevance and necessity to expand the model in two directions. The first direction is

the use of additional, alternative situations which extend beyond the original contingency

categories and predetermined consumer situations used until recently. The second

direction is the addition of other elements beyond the utilitarian and informational

components. This can be translated to further attempts to incorporate more situational

aversive consequences that are meaningful in other environments. Until recently aversive

consequences have been incorporated in terms of the cost of buying and now in terms of

consumer confusion but in alternative settings other aversive elements could be

meaningful.

Regarding the framework of rule-governed behaviour, multiple gains are to be attained if

1) more research treats confusion as a case of track and 2) more constructs, like confusion,

are treated at the level proposed by this study. This treatment based on rule-governed

behaviour could extend their understanding, theoretical implications and allow their

incorporation in more theoretical construals. Specifically for confusion, the application of

the framework of rule-governed behaviour allowed an alternative understanding for the

construct to be developed which facilitated its use in a behavioural model. The

incorporation of confusion in the BPM pointed to the necessity to examine the construct

further and established an alternative view of the relationship of confusion with the

emotional and behavioural responses. The relationship proposed and examined by this

study goes beyond the usual claim that emotional responses are expected to mediate the

relationship between confusion and behavioural responses; the nature of this relationship

is evidently worth of further investigation in its own right.

Further to that, a meaningful philosophical, academic debate on the applications of

intentionality in the study of behaviourism is desirable. This study has indicated some


324

avenues for the application of this framework in the study of the BPM. However as the

framework of intentionality is an evolving topic of interest, meaningful discussion can

take effect, especially regarding the role of intentionality, the role of rule-governed

behaviour and the use of the learning history and consumer situation as starting points for

further theoretical development.

Finally, future research should focus more extensively on the relationship between

confusion and arousal, which seems to carry many levels of theoretical significance.

Based on the finding that measures of confusion do not seem to respond to the measure of

arousal as proposed by the PAD model, it would be relevant to vary both the way

confusion is determined (possibly towards a direction pointed by psychological

research19

) and that of arousal and extend the measurement to other relevant indicators of

arousal like extraversion, neuroticism, sensation-seeking and the Type-A behaviour

pattern (see Furnham, 1984).

Finally, regarding the issue of dominance this study has proposed the possibility that a

further classification of retail situations (pointing to the differences between pleasant and

unpleasant settings) might be required for the more meaningful study of dominance.

10.7. Conclusion and Reflections on the Study

This last section deals with the main personal motivation behind this study along with

publication targets and reflections on the research experience.

The initial motivations behind this thesis have been multiple. A personal concern with the

general topic of consumerism and the implications of our complicated industrial world on

what has been widely called as consumer well-being (although even the most superficial

examination of the relevant literature clearly indicates that no consensus has been reached

19 Refer to p. 116 (chapter 6).


325

on this topic either- see the differing approaches on hedonic and eudemonic wellbeing at

Ryan & Deci, 2001) has been among the first issues that led to the implementation of this

study. Beyond this issue, a further personal, theoretical interest in the formation and

structure of human emotions has turned the interest towards consumer confusion. The idea

of consumer confusion implicates both the elements of consumerism and emotional

ascription and was perceived as one of the areas that allowed for further empirical

inquiries. Finally, the underpinnings of the Behavioural Perspective Model and its

extension towards a novel intentional understanding and way of doing research (Foxall,

2004) offered an unparallel opportunity for exploration.

In order to extend on the aforementioned issue, an examination on the nature of confusion

has revealed several critical issues of interest. Based on an examination of the relevant

psychological research confusion seems to belong to a group of entities that possess both

informational and affective value. It has been examined in the past mainly in terms of its

relative position based on the debate regarding the nature of entities as affective, cognitive

or volitional. In the consumer behaviour literature several conceptualisations of consumer

confusion have been proposed, starting from confusion attributed to the variety of

products and information and their similarity. An issue that is very much discussed in the

relevant literature but lacks conceptual and practical research attempts has been the

implications of confusion for consumers’ emotional and behavioural reactions. In

addition, the literature has reached a point where alternative propositions and debates on

the nature of the construct have not been proposed and the literature seemed to be

reaching a possible stalemate where no new ideas had been introduced or examined.

The other issue of core concern has been the inclusion of cognitive and intentional aspects

to the study of the Behavioural Perspective Model (as described by Foxall, 2004). This

integration and extension took the form of rule-governed behaviour (Foxall, 2013). This


326

study has above all laid the foundations towards a bridging of differing, incommensurable

by some, paradigms of studying human psychology and behaviour. By bringing together

the behavioural and intentional understanding, this is not a way to advocate an

‘epistemological anarchism’ (for example Feyerabend, 1993). Rules for the superior

application of intentional behaviourism have been proposed in previous literature (e.g.

Foxall, 2004; Foxall, 2013). Such rules will be better laid and understood in the future

following the more extensive study of intentional behaviourism. As a result of the

application of these principles to other novel situations and constructs a better

understanding can be achieved. For all these reasons this study is both unique and a

starting point for further research attempts. As always there are some limitations with this

study; when these are faced as opportunities rather than as problematic areas per se, then

the outcome is new interesting prospects for future research. This is why the limitations of

this study were presented along with directions for future research.

On all of these grounds, this study should not be perceived as a final destination but rather

as setting the example for future endeavours. As a result of this research journey, the

following approximate publication targets have been set:

the second theoretical chapter of this thesis is suitable for journals like ‘Emotion’

or ‘Cognition and Emotion’,

the third chapter is ‘a state of the art’ literature review on confusion in consumer

behaviour for journals like ‘Journal of Consumer Behaviour’,

the theoretical contribution of this thesis on the idea of confusion as self-based

rule/ anomy fits a journal like ‘Marketing Theory’,

the empirical findings on confusion and socio-demographic characteristics in a

journal like 'Marketing Intelligence and Planning' and obviously,


327

the empirical incorporation of confusion in the BPM supported by the data in

journals like ‘Journal of Business Research’ and as a chapter in edited books on

consumer behaviour analysis.

As a personal and professional undertaking this study has been painful but worthwhile.

The examination of the relevant literature on emotions and the BPM and identifying

theoretical connections has been a part of this thesis, working with a research agency and

learning from the process of research project management has been another valuable

lesson. Above all else it is the journey that makes the goal meaningful and there is nothing

analogous to a PhD level education to prove that.

REFERENCES

328

11. REFERENCES

Abrahams, R. D. (1986). Ordinary and extraordinary experience, in V.W. Turner and

E.M. Bruner (eds.), The Anthropology of Experience, pp. 45–73. Urbana, IL:

University of Illinois Press.

Abrams, K. M., Courtney, A. M., & Tracy, A. I. (2010). Naturally confused:

Consumers’ perceptions of all natural and organic food products. Agriculture

and Human Values, 27, 365–374.

Abu Hasan, Z. (2011). Interpreting green consumer behaviour: an exploratory

examination of Cardiff consumers. Unpublished PhD Thesis: Cardiff

University.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human

Decision Processes, 50, 179–211.

Alaszewski, A. (2006). Using Diaries for Social Research. Sage: London.

Allison, P. D. (1999). Multiple Regression: A primer. California: Pine Forge Press.

Allison, P. D. (2009). Missing data. In E.R. Millsap and A. Maydeu-Olivares (Eds.)

The Sage Handbook of Quantitative Methods in Psychology. London: Sage

Publications, Ltd.

American Psychological Association (APA) (2013). Glossary of psychological terms.

Available online: <URL: http://www.apa.org/research/action/glossary.aspx#c >

[Accessed February 2013].

Anderson, J. C. (2003). The psychology of doing nothing: Forms of decision avoidance

result from reason and emotion. Psychological Bulletin, 129 (1), 139–167.

Anderson, P. F. (1986). On method in consumer research: a critical relativist

perspective. Journal of Consumer Research, 13 (September), 155–173.

Anonymous, (2013). What is ‘Which’. ‘Which’ official website. Available online:

<URL: http://www.which.co.uk/about-which/who-we-are/overview/>

[Accessed May 2013].

Arnold, M., & Reynolds, K. (2003). Hedonic shopping motivations. Journal of

Retailing, 9 (2), 77–95.

Arnould, J. E., & Price, L. L. (1993). River magic: Extraordinary experience and the

extended service encounter. Journal of Consumer Research, 20, 24–45.

Ashton, C. (1993). A focus on information overload. Managing Service Quality, July,

33–36.

http://www.apa.org/research/action/glossary.aspx#c

http://www.which.co.uk/about-which/who-we-are/overview/

REFERENCES

329

Atkinson, A., & Silverman, D. (1997). Kundera’s immortality: The interview society

and the invention of the self. Qualitative Inquiry, 3 (3), 304–326.

Averill, J. R. (1975). A semantic atlas of emotional concepts. JSAS: Catalogue of

Selected Documents in Psychology, 5 (330).

Aylott, R., & Mitchell, V.-W. (1999). An exploratory study of grocery shopping

stressors. British Food Journal, 101 (9), 683–700.

Babin, J. B., Darden, R. W., & Babin, A. L. (1998). Negative emotions in marketing

research. Affect or artifact? Journal of Business Research, 42, 271–285.

Bagozzi, R. P. (1974). Marketing as an organized behavioural system of exchange.

Journal of Marketing, 38 (October), 77–81.

Bagozzi, R. P., Gopinath, M., & Nyer, P. U. (1999). The role of emotions in marketing.

Journal of Academy of Marketing Science, 27 (2), 184–206.

Baharrel, B., & Denison, D. J. (1995). Involvement in a routine food shopping context.

British Food Journal, 97 (4), 24–29.

Baker, J., Levy, M., & Grewal, D. (1992). An experimental approach to making retail

store environmental decisions. Journal of Retailing, 445–460.

Baker, R. et. al., (2010). Research Synthesis. AAPOR report on online panels. Public

Opinion Quarterly, 1–71.

Balabanis, G., & Craven, S. (1997). Consumer confusion from own brand lookalikes:

An exploratory investigation. Journal of Marketing Management, 13, 299–313.

Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: capturing life as it is lived.

Annual Review of Psychology, 54 (1), 579-616.

Banister, E. M. (2000). Women midlife confusion: “Why am I feeling this way?”

Issues in Mental Health Nursing, 21, 745–764.

Baraldi, A. N., & Enders, C. K. (2010). An introduction to modern missing data

analyses. Journal of School Psychology, 48, 5–37.

Baron, R. M, & Kenny, D. A. (1986). The Moderator-Mediator variable distinction in

social psychological research: Conceptual, strategic, and statistical

considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Barrett, L. F., Mesquita, B., Ochsner, K., N., & Gross, J. J. (2007). The experience of

emotions. Annual Review of Psychology, 58, 373–403.

Baum, W. M. (1995). Rules, culture, and fitness. The Behavior Analyst, 18, 1–21.

Baumeister, R., & Vohs, K. (2003). Willpower, choice, and self-control. In G.

Loewenstein, D. Read and R. Baumeister (Eds.). Time and Decision: Economic

REFERENCES

330

and Psychological Perspectives on Intertemporal Choice. New York: Russell

Sage Foundation, pp. 201–216.

Baumeister, R. F., Sparks, E. A., Stillman, T. F., & Vohs, K. D. (2008). Free will in

consumer behavior: Self-control, ego depletion, and choice. Journal of

Consumer Psychology, 18, 4–13.

Bayus, B. L., & Putsis, Jr. W. P. (1999). Product proliferation: An empirical analysis of

product lines determinants and market outcomes. Marketing Science, 18 (2),

137–153.

Belk, R. (1988). Possessions and the extended self. Journal of Consumer Research, 15

(September), 139–168.

Belk, R. W., Wallendorf, M., & Sherry, J. F. (1989). The sacred and the profane in

consumer behaviour: theodicy on the odyssey. Journal of Consumer Research,

16 (June), 1–37.

Bellenger, D. N., & Korgaonkar, P. K. (1990). Profiling the recreational shopper.

Journal of Retailing, 66 (Fall), 77–82.

Bennett, M. R., & Hacker, P. M. S. (2003). Philosophical Foundations of

Neuroscience. Oxford: Blackwell.

Bennett, M. R., & Hacker, P. M. S. (2005). Emotion and cortical-subcortical function:

conceptual developments. Progress in Neurobiology, 75, 29–52.

Benton, T., & Craib, I. (2001). Philosophy of social science: The philosophical

foundations of social thought. New York: Palgrave.

Bergvist, L., & Rossiter, J. R. (2007). The predictive validity of multiple-item versus

single-item measures of the same constructs. Journal of Marketing Research,

44 (May), 175–184.

Berlyne, D. E. (1960). Conflict, Arousal and Curiosity. New York: McGraw Hill.

Bettman, J. R. (1979). An Information Processing Theory of Consumer Choice.

Reading, MA: Addison-Wesley.

Bettman, J. R., Eric, J. J., & Payne, W. J. (1990). A componential analysis of cognitive

effort in choice. Organizational Behaviour and Human Decision Processes, 45

(February), 111–139.

Bitner, M. J. (1992). Servicescapes: the impact of physical surroundings on customers

and employees. Journal of Marketing, 56 (April), 57–71.

Blackman, D. E. (1974). Operant conditioning: An experimental analysis of behaviour.

London: Methuen & Co Ltd.

Blaikie, N. (2000). Designing Social Research. Cambridge: Polity Press.

REFERENCES

331

Blakely, E., & Schlinger, H. (1987). Rules: Function-altering contingency-specifying

stimuli. The Behavior Analyst, 10 (2), 183–187.

Borsboom, D. (2005). Measuring the mind: Conceptual issues in contemporary

psychometrics. Cambridge: Cambridge University Press.

Bougie, R., Pieters, R., & Zeelenberg, M. (2003). Angry customers don’t come back,

they get back: The experience and behavioral implications of anger and

dissatisfaction in services. Journal of the Academy of Marketing Science, 31

(4), 377–393.

Brace, N., Kemp, R., & Snelgar, R. (2012). SPSS for psychologists: a guide to data

analysis using SPSS for Windows (5th Edition). New York: Palgrave

Macmillan.

Brehm, J. W. (1956). Post- decision changes in the desirability of alternatives. Journal

of Abnormal and Social Psychology, 52, 384–389.

Brehm, J. W. (1972). Responses to the loss of freedom: A theory of psychological

reactance. Morristown, NJ: General Learning Press.

Briggs, C. (1986) Learning How to Ask: A Sociolinguistic Appraisal of the Role of the

Interview in Social Science Research. Cambridge: Cambridge University Press.

Broniarczyk, S. M. (2008). Product assortment. In Handbook of Consumer Psychology.

In Haugtvedt, C. P., Herr, P. M. and Kardes F. R. (Eds.). New York: Taylor and

Francis Group.

Broniarczyk, S. M., Hoyer, W. D., & McAlister, L. (1998). Consumers’ perceptions of

the assortment offered in a grocery category: The impact of item reduction.

Journal of Marketing Research, 35, 166–176.

Brownstein, A. J., & Shull, R. L. (1985). A rule for the use of the term, "Rule-governed

behavior". The Behavior Analyst, 8, 265–267.

Bryman, A., & Bell, E. (2007). Business Research Methods. Oxford: University Press.

Burns, A. C., & Bush, R. F. (2010). Marketing Research (6th Edition). New Jersey;

Pearson Education Inc.

Burns, R. B. (2010). Introduction to Research Methods (4th

Edition). London: Sage.

Burt, S., & Davis, S. (1999). Follow my leader? Lookalike retailer brands in non-

manufacturer-dominated product markets in the UK. International Review of

Retail and Distribution Consumer Research, 9 (2), 163–185.

Cahill, D. J. (1995). We sure as hell confuse ourselves, but what about the customers?

Marketing Intelligence & Planning, 13 (4), 5–9.

http://en.wikipedia.org/wiki/Measuring_the_Mind

http://en.wikipedia.org/wiki/Cambridge_University_Press

REFERENCES

332

Calder, B. J., Phillips, L. W., & Tybout, A. M. (1982). The concept of external validity.

Journal of Consumer Research, 9, 240–244.

Carmon, Z., Wertenbroch K., & Zeelenberg, M. (2003). Option attachment: When

deliberating makes choosing feel like losing. Journal of Consumer Research, 30

(1), 15–29.

Carù, A., & Cova, B. (2003). Revisiting consumption experience; a more humble but

complete view of the concept. Marketing Theory, 3 (2), 267–286.

Carver, C. S., & Scheier, M. F. (1994). Situational coping and coping dispositions in a

stressful transaction. Journal of Personality and Social Psychology, 66 (1),

184–195.

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies:

A theoretically based approach. Journal of Personality and Social Psychology,

56 (2), 267–283.

Casini, L., Cavicchi, A., & Corsi, A. M. (2008). Trends in the British wine market and

consumer confusion. British Food Journal, 110 (6), 545–558.

Catania, A. C. (1986). On the difference between Verbal and Nonverbal Behaviour.

The Analysis of Verbal Behavior, 4, 2–9.

Catania, A. C. (1989). Rules as classes of verbal behavior: A reply to Glenn. The

Analysis of Verbal Behavior, 7, 49–50.

Catania, A. C. (1998). Learning (4th Edition). Upper Saddle River, NJ: Prentice Hall.

Catania, A. C., Lowe, C. F. &, Horne, P. (1990). Nonverbal behaviour correlated with

the shaped verbal behaviour of children. The Analysis of Verbal Behavior, 8,

43–55.

Cerutti, D. T. (1989). Discrimination theory of rule-governed behavior. Journal of the

Experimental Analysis of Behavior, 51, 259–276.

Charmaz, K. (2006) Constructing Grounded Theory: A Practical Guide Through

Qualitative Analysis. London: Sage.

Chernev, A. (2003a). When more is less and less is more: The role of ideal point

availability and assortment in consumer choice. Journal of Consumer Research,

30, 170–184.

Chernev, A. (2003b). Product assortment and individual decision processes. Journal of

Personality and Social Psychology, 85, 151–162.

Chernev, A. (2006). Decision focus and consumer choice among assortments. Journal

of Consumer Research, 33, 50–59.

REFERENCES

333

Chryssochoidis, G. (2000). Repercussions of consumer confusion for late introduced

differentiated products. European Journal of Marketing, 34 (5/6), 705–722.

Churchill, G. A., Jr. (1979). A paradigm for developing better measures of marketing

constructs. Journal of Marketing Research, 16 (February), 64–73.

Clore, G. L. (1992). Cognitive phenomenology: Feelings and the construction of

judgement. In L.L. Martin & A. Tesser (Eds.). The Construction of Social

Judgements (pp. 133–163). Hillsdale, NJ:Erlbaum.

Clore, G. L., Ortony, A., & Foss, M. A. (1987). The psychological foundations of the

affective lexicon. Journal of Personality and Social Psychology, 53, 751–766.

Clore, G. L., Schwarz, N., & Conway, M. (1994). Affective causes and consequences

of social information processing. In Handbook of Social Cognition, Vol. 1,

Robert S. Wyer and Thomas K. Srull, (Eds). Hillsdale, NJ: Lawrence Erlbaum

Associates, 323–417.

Cook, T. D., & Campbell, D. T. (1979). Quasi-Experimentation: Design and analysis

for field settings. Boston: Houghton Mifflin.

Cooper, J. (2007). Cognitive dissonance: 50 years of a classic theory. London: Sage

Publications.

Cortina, J.M. (1993). What is coefficient alpha? An examination of theory and

applications. Journal of Applied Psychology, 78 (1), 98–104.

Crane, T. (1998). Intentionality as the mark of the mental. In A. O’Hear (Ed.)

Contemporary Issues in the Philosophy of Mind. Cambridge: Cambridge

University Press, 229–251.

Crane, T. (2003). The intentional structure of consciousness. In A. Jokic and Q. Smith

(Eds.) Consciousness: New Philosophical Perspectives. Oxford and New York:

Oxford University Press.

Crane, T. (2007). Intentionalism. In B. McLaughlin & A. Beckermann (Eds.) Oxford

Handbook to the Philosophy of Mind (pp. 474-493). New York: Oxford

University Press.

Creswell, J. W, & Plano Clark, V. L. (2007). Designing and Conducting Mixed

Methods Research (3rd

Edition). California: Sage Publications.

Cross, G. S. (2000). An All-Consuming Century: Why Commercialism Won in Modern

America. New York: Columbia University Press.

Cupchik, G. (2001). Constructivist realism: An ontology that encompasses positivist

and constructivist approaches to the social sciences. Forum Qualitative

Sozialforschung/ Forum: Qualitative Social Research, 2(1), Art. 7. Available

online: < URL: http://nbnresolving.de/urn:nbn:de:0114-fqs010177>

http://nbnresolving.de/urn:nbn:de:0114-fqs010177

REFERENCES

334

Data Protection Act (1998). Reprinted incorporating corrections (2000). The

Stationery Office: Norwich. Available online: <URL:

http://www.legislation.gov.uk/ukpga/1998/29/introduction> [Accessed June

2012].

Damasio, A. (2005). Descartes’ error: Emotion, reason, and the human brain. New

York: Penguin.

Darwin, C. (1965). The Expression of Emotions in Man and Animals. Chicago:

Chicago University Press. (Original work published 1872).

Davies, J. D., Foxall, G. R., & Pallister, J. G. (2002). Beyond the intention-behaviour

mythology: An integrated model of recycling. Marketing Theory, 2, 29–113.

Davies, A., & Fitchett, A. J. (2005). Beyond incommensurability? Empirical expansion

on diversity in research. European Journal of Marketing, 39 (3), 272–293.

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance

of information technology. MIS Quarterly, 13 (3), 319–340.

Dawson, S., Bloch, P. H., & Ridgway, N. M. (1990). Shopping motives, emotional

states, and retail outcomes. Journal of Retailing, 66 (4), 408.

Deci, E. L. (1975). Intrinsic Motivation. New York: Plenum Press.

Deci, E. L., & Ryan, R. M. (1985). Intrinsic Motivation and Self-Determination in

Human Behavior. New York: Plenum Press.

Denegri-Knott, J., Zwick, D., & Schroeder, E. J. (2006). Mapping consumer power: an

integrative framework for marketing and consumer research. European Journal

of Marketing, 40 (9/10), 950–971.

Dennett, D. C. (1983). Intentional systems in cognitive ethology: The ‘Panglosian’

paradigm defended. The Behavioral and Brain Sciences, 6, 343–339.

Dennett, D. C. (1987). The Intentional Stance (1st Printing). Cambridge,

Massachusetts: The MTI Press.

Dennett, D. C. (2007). Intentional Systems Theory. In B. McLaughlin & A.

Beckermann (Eds.) Oxford Handbook to the Philosophy of Mind (pp. 339-350).

New York: Oxford University Press.

Dennis, J. M. (2001). Are internet panels creating professional respondents? Marketing

Research, 13 (2), 34–38.

De Rosa, A., S. (1993). Social representations and attitudes: problems of coherence

between the theoretical definition and procedure of research. Papers on Social

Representations, 2 (3), 178–192.

http://www.legislation.gov.uk/ukpga/1998/29/introduction

REFERENCES

335

Deshpande, R. (1983). Paradigms lost: on theory and method in research in marketing.

Journal of Marketing, 47 (Fall), 101–110.

De Vaus, D. (2002). Surveys in Social Research (5th

Edition.). London: UCLPress and

Routledge.

Dhar, R. (1997). Consumer preference for a no-choice option. Journal of Consumer

Research, 24 (September), 215–231.

Diamantopoulos, A., Schlegelmich, B.B., Sinkovics, R.R., & Bohlen, G. M. (2003).

Can socio-demographics still play a role in profiling green consumers? A

review of the evident and an empirical investigation. Journal of Business

Research, 56, 465–480.

DiClemente, D. F. & Hantula, D. A. 2003. Applied behavioral economics and

consumer choice. Journal of Economic Psychology, 24 (5), 589–602.

Donovan, R. J., & Rossiter, J. R. (1982). Store atmosphere: An environment

psychology approach. Journal of Retailing, 58 (Spring), 34–57.

Donovan, R. J., Rossiter, J. R., Marcoolyn, G., & Nesdale, A. (1994). Store atmosphere

and purchasing behavior. Journal of Retailing, 70 (3), 283–294.

Dowling, G., & Staelin, R. (1994). A model of perceived risk and intended risk

handling activity. Journal of Consumer Research, 21 (1), 119–135.

Drolet, A. L., & Morrison, D. G. (2001). Do we really need multiple-item measures in

service research? Journal of Service Research, 3 (3), 196-204.

Drummond, G. (2004). Consumer confusion: reduction strategies in higher education.

International Journal of Education Management, 18 (5), 317–323.

Drummond, G., & Rule, G. (2005). Consumer confusion in the UK wine industry.

Journal of Wine Research, 16 (1), 55–64.

Dube L., & Morgan, M. S. (1998). Capturing the dynamics of in-process consumption

emotions and satisfaction in extended service transactions. International

Journal of Research in Marketing, 15, 309–320.

Dwyer, F. R., Shurr, P. H., & Oh, S. (1987). Developing buyer and seller relationships.

Journal of Marketing, 51 (April), 11–27.

Easley, W. R., Madden, S. C., & Dunn, G. M. (2000). Conducting marketing science:

The role of replication in the research process. Journal of Business Research,

48, 83–92.

Eagly, A. H. & Chaiken, S. (1993). The Psychology of Attitudes. Harcourt Brace

Jovanovich, Fort Worth: Orlando, FL.

REFERENCES

336

Edell, J. A., & Burke, M. C. (1987). The power of feelings in understanding

advertising effects. Journal of Consumer Research, 14 (4), 421–433.

Ekman, P., & Friesen, W. V. (1971). Constants across cultures in the face and

emotions. Journal of Personality and Social Psychology, 17, 124–129.

Ekman, P. (1992). An argument for basic emotions. Cognition and Emotion, 6, 169–

200.

Ekman, P. (2001). Facial Expressions. In: C. Blakemore & S. Jennett (Eds.) Oxford

Companion to the Body. London: Oxford University Press.

Ekman, P. (2003). Facial Expressions of Emotion. In R., J., Davidson, K., R., Scherer,

& H., H., Goldsmith (Eds.) Handbook of Affective Sciences, pp. 415–431. New

York: Oxford University Press.

Elliot, R., & Ritson, M. (1995). Practicing existential consumption: The lived meaning

of sexuality in advertising. Advances in Consumer Research, 22, 740–745.

Elliott, M. T., & Speck, P. S. (1998). Consumer perceptions of advertising clutter and

its impact across various media. Journal of Advertising Research, 38 (1), 29–

41.

Ellsworth, G. P. (2003). Confusion, concentration and other emotions of interest:

commentary on Rozin and Cohen (2003). Emotion, 3 (1), 81–85.

Ellsworth, G. P., & Scherer, R. K. (2003). Appraisal processes in emotion. In J. R.,

Davidson, R. K., Scherer, & H. H., Goldsmith (Eds.) Handbook of Affective

Sciences, pp. 572–599. New York: Oxford University Press.

Eppler, J. M., & Mengis, J. (2004). The concept of information overload: A review of

literature from organization science, accounting, marketing, MIS, and related

disciplines. The Information Society, 20, 325–344.

ESRC. 2010. Research Ethics Framework, available online: <URL:

http://www.esrc.ac.uk/_images/Framework_for_Research_Ethics_tcm8-

4586.pdf> [Accessed June 2012].

Ess, C., & AoIR ethics working committee, (2002). Ethical decision-making and

Internet research: Recommendations from the AoIR ethics working committee.

Approved by AoIR, November 27, 2002. Available online: <URL:

www.aoir.org/reports/ethics.pdf> [Accessed July 2012].

Eynon, R., Fry, J., & Schroeder, R. (2008). The ethics of internet research. In N. G.,

Fielding, R. M, Lee, & G., Blank (Eds.) The Sage Handbook of Online

Research Methods, pp. 23–41. London: SAGE Publications Ltd.

Faison, E. W. J. (1977). The neglected variety drive. Journal of Consumer Research, 4

(December), 172–175.

http://www.esrc.ac.uk/_images/Framework_for_Research_Ethics_tcm8-4586.pdf

http://www.esrc.ac.uk/_images/Framework_for_Research_Ethics_tcm8-4586.pdf

http://www.aoir.org/reports/ethics.pdf

REFERENCES

337

Fasolo, B., McClelland, G. H., & Todd, P. M. (2007). Escaping the tyranny of choice:

When fewer attributes make choice easier. Marketing Theory, 7 (1), 13–26.

Fasolo, B., Hertwig, R., Huber, M., & Ludwig, M. (2009). Size, entropy and density:

What is the difference that makes the difference between small and large real-

world assortment? Psychology and Marketing, 26 (3), 254–279.

Fehr, B., & Russell, A. J. (1984). Concept of emotion viewed from a prototype

perspective. Journal of Experimental Psychology General, 113 (3), 464–486.

Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford

University Press.

Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced

compliance. Journal of Abnormal and Social Psychology, 58, 203–210.

Feyerabend, P. (1993). Against method (3rd

Edition). London: Verso.

Firat, A., & Venkatesh, A. (1995). Liberatory postmodernism and the re-enchantment

of consumption. Journal of Consumer Research, 22 (3), 239–267.

Fishbein, M. & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An

introduction to theory and research. Reading, MA: Addison-Wesley.

Foxall, G. R. (1986). Theoretical progress in consumer psychology: the contribution of

a behavioural analysis of choice. Journal of Economic Psychology, 7, 393–414.

Foxall, G. R. (1987). Radical behaviourism and consumer research. Theoretical

promise and empirical problems. International Journal of Research in

Marketing, 4, 111–129.

Foxall, G.R. (1990). Consumer Psychology in Behavioural Perspective. London and

New York: Routledge.

Foxall, G. R. (1992a). The behavioural perspective model of purchase and

consumption: from consumer theory to marketing practice. Journal of the

Academy of Marketing Science, 20, 189–198.

Foxall, G. R. (1992b). The consumer situation: An integrative model for research in

marketing. Journal of Marketing Management, 8, 383–404.

Foxall, G. R. (1993). Consumer behaviour as an evolutionary process. European

Journal of Marketing, 27 (8), 46–58.

Foxall, G. R. (1995). Science and interpretation in consumer research: a radical

behaviourist perspective. European Journal of Marketing, 29 (9), 3–99.

Foxall, G. R. (1996). Consumers in context. The BPM research program. London and

New York: Routledge.

REFERENCES

338

Foxall, G. R. (1997a). Marketing Psychology: The Paradigm in the Wings. London:

MacMillan Press LTD.

Foxall, G. R. (1997b). The emotional texture of consumer environments: a systematic

approach to atmospherics. Journal of Economic Psychology, 18 (5), 505–523.

Foxall, G. R. (1997c). Affective responses to consumer situations. The International

Review of Retail, Distribution and Consumer Research, 7 (3), 191–225.

Foxall, G. R. (1998). Radical behaviourist interpretation: generating and evaluating an

account of consumer behavior. The Behavior Analyst, 21, 321–354.

Foxall, G. R. (1999a). The Behavioural Perspective Model: consensibility and

consensuality. European Journal of Marketing, 33 (5/6), 570–596.

Foxall, G. R. (1999b). The contextual stance. Philosophical Psychology, 12 (1), 25–46.

Foxall, G. R. (2000). The contextual stance in consumer research. European Journal of

Marketing, 34 (7), 768–779

Foxall, G. R. (2001). Foundations of consumer behaviour analysis. Marketing Theory,

1 (2), 165–199.

Foxall, G. R. (2004). Context and Cognition. Interpreting Complex Behavior. Reno,

NV: Context Press.

Foxall, G. R. (2007a). Explaining consumer choice: coming to terms with

intentionality. Behavioural Processes, 75 (2), 129–145.

Foxall, G. R. (2007b). Intentional Behaviorism. Behavior and Philosophy, 35, 1–55.

Foxall, G. R. (2007c). Explaining Consumer Choice. London: Palgrave.

Foxall, G. R. (2008). Intentional Behaviorism Revisited. Behavior and Philosophy, 36,

113–155.

Foxall, G. R. (2010a). Interpreting Consumer Choice: The Behavioural Perspective

Model. New York: Routledge.

Foxall, G. R. (2010b). Accounting for consumer choice: Interpersonal decision-making

in behavioural perspective. Marketing Theory, 10 (4), 315–345.

Foxall, G. R. (2011). Brain, emotion and contingency in the explanation of consumer

behaviour published in G. P. Hodgkinson & J. K. Ford (Eds.) (2011).

International Review of Industrial and Organisational Psychology. 26. pp. 47–

92. Chichester, UK: Wiley.

Foxall, G. R. (2013). Intentionality, symbol, and situation in the interpretation of

consumer choice. Marketing Theory, 13 (1), 105–127.

REFERENCES

339

Foxall, G. R., & Greenley, G. E. (1998). The affective structure of consumer situations.

Environment and Behavior, 30, 781–798.

Foxall, G. R., & Greenley, G. E. (1999). Consumers‘ emotional responses to service

environments. Journal of Business Research, 46 (2), 149–158.

Foxall, G. R., & Greenley, G. E. (2000). Predicting and explaining responses to

consumer environments: an empirical test and theoretical extension of the

Behavioural Perspective Model. The Service Industries Journal, 20 (2), 39–63.

Foxall, G. R., & Soriano, M. Y. (2005). Situational influences on consumers' attitudes

and behavior. Journal of Business Research, 58 (4), 518–525.

Foxall, G. R., & Soriano, M. Y. (2011). Influence of reinforcement contingencies and

cognitive styles on affective responses: An examination of Roll’s theory of

emotion in the context of consumer choice. Journal of Applied Social

Psychology, 41 (10), 2508–2537.

Foxall, G. R., Oliveira-Castro, J. M., James, V. K., Soriano, M. Y., & Sigurdsson, V.

(2006). Consumer behavior analysis and social marketing: The case of

environmental conservation. Behavior and Social Issues, 15, 101–124.

Foxall, G. R., Soriano, M. Y., Yousafzai, S. Y., & Javed, U. (2012). The role of

neurophysiology, emotion and contingency in explanation of consumer choice.

In V. Wells & G. R., Foxall (Eds) (2012). Handbook of Developments in

Consumer Behaviour. pp. 461–522. Cheltenham: Edward Elgar Publishing

Limited.

Foxall, G. R., Goldsmith, R. E., & Brown, S. (1998). Consumer psychology for

marketing. London: International Thomson Business Press.

Foxman, E. R., Berger, P. W., & Cote, J. A. (1992). Consumer brand confusion: A

conceptual framework. Psychology & Marketing, 9 (2), 123–141.

Foxman, E. R., Muehling, D. D., & Berger, P. W. (1990). An investigation of factors

contributing to consumer brand confusion. Journal of Consumer Affairs, 24 (1),

170–189.

Friedman, M. P. (1966). Consumer confusion in the selection of super-market

products. Journal of Applied Psychology, 50 (6), 529–534.

Frijda, N. H, Kuipers, P., & Ter Schure, E. (1989). Relations among emotion,

appraisal, and emotional action readiness. Journal of Personality and Social

Psychology, 57 (2), 212–228.

Frijda, N. H. (2008). The psychologist’s point of view. In M. Lewis and J.M. Haviland-

Jones (ed.) Handbook of Emotions, 3rd ed., pp. 68–87. New York: Guilford.

REFERENCES

340

Fuchs, C., & Diamantopoulos, A. (2009). Using single-item measures for construct

measurement in management research. Conceptual issues and application

guidelines. DBW, 69 (2), 195–210.

Furnham, A. (1984). Extraversion, sensation-seeking, stimulus screening and type A-

behaviour pattern: The relationship between various measures of arousal.

Personality and Individual Differences, 5 (2), 133–140.

Galizio, M. (1979). Contingency-shaped and rule-governed behavior: Instructional

control of human loss avoidance. Journal of the Experimental Analysis of

Behavior, 31, 53–70.

Gardner, M. P. (1985). Mood states and consumer behaviour: A critical review.

Journal of Consumer Research, 12 (December), 281–300.

Gillham, B. (2000). The Research Interview. London & New York: Continuum.

Greenleaf, E. A., & Lehmann, D. R. (1995). Reasons for substantial delay in consumer

decision-making. Journal of Consumer Research, 22 (September), 186–199.

Glenn, S. S. (1987). Rules as environmental events. The Analysis of Verbal Behavior,

5, 29–32.

Grimes, A, & Kitchen, J. P. (2007). Researching mere exposure effects to advertising.

International Journal of Market Research, 49 (2), 191–221.

Golodner, F. L. (1993). Healthy confusion for consumers. Journal of Public Policy and

Marketing, 12 (1), 130–134.

Göritz, A. S. (2004). The impact of material incentives on response quantity, response

quality, sample composition, survey outcome, and cost in online access panels.

International Journal of Market Research, 46 (3), 327–345.

Göritz, A. S. (2007). Using online panels in psychological research. In A. Joinson, K.

McKenna, T. Postmes and U. D. Reips (eds.), pp.473–495. The Oxford

Handbook of Internet Psychology, Oxford: Oxford University Press.

Göritz, A. S., Reinhold, N., & Batinic, B. (2000) Online panels. In B. Batinic, U. - D.

Reips, M. Bosnjak and A. Werner (eds.), Online Social Sciences, pp. 29–52.

Gorn, G. J. (1982). The effects of music in advertising on choice behavior: A classical

conditioning approach. Journal of Marketing, 46, 94–101.

Gourville, J. T., & Soman, D. (2005). Overchoice and assortment type: When and why

variety backfires. Marketing Science, 24 (3), 382–395.

Goulding, C. (1999). Consumer research, interpretive paradigms and methodological

ambiguities. European Journal of Marketing, 33 (9/10), 859–873.

REFERENCES

341

Goulding, C. (2005). Grounded theory, ethnography and phenomenology: a

comparative analysis of three qualitative strategies for marketing research.

European Journal of Marketing, 39 (3/4), 294–308.

Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research, in

N. K. Denzin, & Y. S. Lincoln (Eds.), Handbook of Qualitative Research,

Newbury Park, CA: Sage Publications, pp. 105–117.

Guba, E. G., & Lincoln Y. S. (1998) Competing paradigms in qualitative research. In

N. K. Denzin and Y.S., Lincoln (Eds.), The Landscape of Qualitative Research,

CA: Sage Thousand Oaks, pp. 195–222.

Guba, E. G., & Lincoln, Y. S. (2005). Paradigmatic controversies, contradictions, and

emerging influences, in N. K. Denzin, & Y. S. Lincoln (Eds.), The Sage

Handbook of Qualitative Research (3rd Edition), pp. 191–215. Thousand Oaks,

CA: Sage.

Guiry, M., Mägi, W. A., & Lutz, J., R. (2006). Defining and measuring recreational

shopper identity. Journal of the Academy of Marketing Science, 34 (1), 74–83.

Hacker, P. M. S. (2004). The conceptual framework for the investigation of emotions.

International Review of Psychiatry, 16 (3), 199–208.

Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data

analysis (5th Edition), Saddle River, New Jersey: Prentice Hall, Inc.

Hasler, C. M. (2008). Health claims in the United States: An aid to the public or a state

of confusion? The Journal of Nutrition, 1216S–1220S.

Havlena, J. W., & Holbrook, B. M. (1986). The varieties of consumption experiences:

comparing two typologies of emotion in consumer behaviour. Journal of

Consumer Research, 13, 394–404.

Havlena, J. W., Holbrook, B. M., & Lehmann, R. D. (1989). Assessing the validity of

emotional typologies. Psychology & Marketing, 6, 97–112.

Healy, M., & Perry, C. (2000). Comprehensive criteria to judge validity and reliability

of qualitative research within the realism paradigm. Qualitative Market

Research: An International Journal, 3 (3), 118–126.

Heider, F. (1958). The Psychology of Interpersonal Relations. New York: Wiley.

Henryks, J., & Pearson, D. (2010). Misreading between the lines: Consumer confusion

over organic food labelling. Australian Journal of Communication, 37 (3), 73–

86.

Hess, U. (2003). Now you see it, now you don’t—The confusing case of confusion as

an emotion: Commentary on Rozin and Cohen (2003). Emotion, 3 (1), 76–80.

REFERENCES

342

Hirschman, O. A. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms,

Organizations, and States. Cambridge, MA: Harvard University Press.

Hirschman, E. C. (1986). Humanistic enquiry in marketing research: philosophy,

method and criteria. Journal of Marketing Research, 23 (August), 327–349.

Hirschman, E. C., & Holbrook, M. B. (1982). Hedonic consumption: Emerging

concepts, methods and propositions. Journal of Marketing, 46 (summer), 92–

101.

Ho, R. (2006). Handbook of univariate and multivariate data analysis and

interpretation with SPSS. Florida: Chapman and Hall/CRC.

Hoch, S., Bradlow, E. T., & Wansink, B. (2002). The variety of an assortment.

Marketing Science, 18 (4), 527–546.

Holbrook, M. B. (2000). The Millennial consumer in the texts of our times: Experience

and entertainment. Journal of Macromarketing, 20 (2), 178–192.

Holbrook, M. B., & Hirschman, E. C. (1982). The experiential aspects of consumption:

Consumer fantasy, feelings and fun. Journal of Consumer Research, 9 (2), 132–

140.

Howell, C. D. (2007). Statistical Methods for Psychology (6th

Edition). Belmont, USA:

Thomson Wadsworth (part of the Thomson Corporation).

Huang, M., H. (2001). The theory of emotions in marketing. Journal of Business and

Psychology, 16 (2), 239–247.

Hubbard, R., & Armstrong, J. S. (1994). Replications and extensions in marketing-

Rarely published but quite contrary. International Journal of Research in

Marketing, 11, 233–248.

Huffman, C., & Kahn, E. B. (1998). Variety for sale: Mass customization or mass

confusion? Journal of Retailing, 74 (4), 491–513.

Hunt, D. S. (1991). Positivism and paradigm dominance in consumer research: toward

critical pluralism and rapprochement. Journal of Consumer Research, 18, 32–

44.

Hwang, M. I., & Lin, J. W. (1999). Information dimension, information overload and

decision quality. Journal of Information Science, 25, 213–318.

Hyde, F. K. (2000). Recognising deductive processes in qualitative research.

Qualitative Market Research: An International Journal, 3 (2), 82–90.

Ippolito, P. M., & Mathios, A. D. (1994). Nutrition information and policy: A study of

U.S. food production trends. Journal of Consumer Policy, 17 (3), 271–305.

http://en.wikipedia.org/wiki/Albert_O._Hirschman

REFERENCES

343

Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire

too much of a good thing? Journal of Personality and Social Psychology, 79

(6), 995–1006.

Izard, C. E. (1994). Innate and universal facial expressions: Evidence from

developmental and cross-cultural research. Psychological Bulletin, 115, 288–

299.

Izard, C. E. (2007). Basic emotions, natural kinds, emotions schemas, and a new

paradigm. Perspectives on Psychological Science, 2 (3), 260–280.

Jacoby, J. (1977). Information load and decision quality: Some contested issues.

Journal of Marketing Research, 14 (4), 569–573.

Jacoby, J. (1984). Perspectives on information overload. Journal of Consumer

Research, 10 (4), 432–435.

Jacoby, J., Speller, D. E., & Kohn, C. A. (1974a). Brand choice behaviour as a function

of information load. Journal of Marketing Research, 11 (1), 63–69.

Jacoby, J., Speller, D. E., & Kohn, C. A. (1974b). Brand choice behaviour as a function

of information load: Replication and extension. The Journal of Consumer

Research, 1, 33–42.

Jacoby, J., & Maureen, M. (1998). ‘Not manufactured or authorized by...’: recent

federal cases involving trademark disclaimers. Journal of Public Policy &

Marketing, 17, 97–108.

Kahn, B. E. (1995). Consumer variety-seeking among goods and services: An

integrative review. Journal of Retailing and Consumer Services, 2 (3), 139–

148.

Kahn, B. E. (1998). Dynamic relationships with customers: high-variety strategies.

Journal of the Academy of Marketing Science, 26 (1), 45–53.

Kahn, B. E., & Ratner, R. (2005). Variety for the sake of variety? Diversification

motives in consumer choice. In Ratneshwar, S. and Mick, D. G. (Eds). Inside

Consumption: Frontiers of Research on Consumer Motives, Goals, and

Desires. London: Routledge.

Kaltcheva, V, & Weitz, B. (2006). When should a retailer create an exciting store

environment? Journal of Marketing, 70 (1), 107–118.

Kapferer, J. -N. (1995). Brand confusion: empirical study of a legal concept.

Psychology and Marketing, 12, 551–568.

Kasper, H., Bloemer, J., & Driessen, P. H. (2010).Coping with confusion: The case of

the Dutch mobile phone market. Managing Service Quality, 20 (2), 140–160.

REFERENCES

344

Keinan, G. (1987). Decision making under stress: Scanning of alternatives under

controllable and uncontrollable threats. Journal of Personality and Social

Psychology, 52, 639–644.

Kellaris, J. J., & Kent, R. J. (1993). An exploratory investigation of responses elicited

by music varying in tempo, tonality, and texture. Journal of Consumer

Psychology, 2 (October), 381–401.

Keller, K. L., & Staelin, R. (1987). Effects of quality and quantity of information on

decision effectiveness. Journal of Consumer Research, 14 (September), 200–

213.

Kelley, H. H., & Michela, J. L. (1980). Attribution theory and research. Annual Review

of Psychology, 31, 457–501.

Keltner, D., & Shiota, M. N. (2003). New displays and new emotions: A commentary

on Rozin and Cohen (2003). Emotion, 3, 86–91.

Kent, R. (2007). Marketing Research Approaches, Methods and Applications in

Europe. London: Thomson Learning.

Khermouch, G. (1994). Packard Bell becomes contrarian with launch of sub-brand:

Spectria. Brandweek, 35 (25), 9.

Kuhn, T. (1962). The Structure of Scientific Revolutions. Chicago University Press,

Chicago, IL.

Kim, Y. K. (2001). Experiential retailing: An interdisciplinary approach to success in

domestic and international retailing. Journal of Retailing and Consumer

Services, 8, 287–289.

Kinnear, P. R., & Gray, C. D. (2000). SPSS for windows made simple. East Sussex;

Psychology Press, Ltd.

Kleinginna, P. R., & Kleinginna, A. M. (1981). A categorized list of emotion

definitions with suggestions for a consensual definition. Motivation and

Emotion, 5, 345–379.

Knight, F.H. (1948). Risk, uncertainty and profit. Houghton-Mifflin, Boston, MA.

Knobe, J. (2003). Intentional action in folk psychology: An experimental investigation.

Philosophical Psychology, 16, 309–324.

Kohli, A. K., & Jaworski, B. J. (1990). Market orientation: The construct, research

propositions, and managerial implications. The Journal of Marketing, 54 (2), 1–

18.

Kotler, P. (1973). Atmosphere as a marketing tool. Journal of Retailing, 49 (Winter),

48–64.

REFERENCES

345

Kotler, P, & Armstrong, G. (2008). Principles of Marketing. Pearson Education: New

Jersey.

Krauss, S. E. (2005). Research paradigms and meaning making: a primer. The

Qualitative Report, 10 (4), 758–770.

Kuhn, T. (1970). The Structure of Scientific Revolutions. Chicago: Chicago University

Press.

Kunkel, H. J. (1997). The analysis of rule-governed behavior in social psychology.

Psychological Record, 47 (4), 698–716.

Lancaster, K. (1990). The economics of product variety: A survey. Marketing Science,

9 (3), 189–206.

Lane, R. E. (2000). The Loss of Happiness in Market Economies. New Haven, CT:

Yale University Press.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). Emotion, motivation, and

anxiety: brain mechanisms and psychophysiology. Society of Biological

Psychiatry, 44, 1248–1263.

Laroche, M., Cleveland, M. Bergeron, J. &, Goutaland, C., (2003). The knowledge

experience-evaluation relationship: A structural equations modelling test of

gender differences. Canadian Journal of Administrative Sciences, 20 (3), 246–

259.

Laros, F. J. M., & Steenkamp, J.-B. E. M. (2005). Emotions in consumer behaviour: A

hierarchical approach. Journal of Business Research, 58, 1437–1445.

Larry, A. & Michael, M. (2012). Deontological ethics, The Stanford Encyclopedia of

Philosophy, in Edward N. Zalta (ed.), URL:

<http://plato.stanford.edu/archives/win2012/entries/ethics-deontological/>.

[Accessed March 2013].

Lazarus, R. S. (1984). On the primacy of cognition. American Psychologist, 39, 124–

129.

Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.

Lazarus, R. S. (1995). Vexing research problems inherent in cognitive- mediational

theories of emotions and some solutions. Psychological Inquiry, 6, 183-196.

Lazarus, R. S., & Folkman, S. (1984). Stress, Appraisal and Coping. New York:

Springer.

Lazarus, R.S., Averill, J.R., & Opton, E.M. (1970). Towards a cognitive theory of

emotion. In M. B. Arnold (Eds.), Feelings and emotions (pp. 207-232). New

York: Academic Press.

REFERENCES

346

LeDoux, J. E. (1996). The emotional brain: The mysterious underpinnings of emotional

life. New York: Simon & Schuster.

Lee, S., Ha, S., & Widdows, R. (2011). Consumer responses to high technology

products: Product attributes, cognitions and emotions. Journal of Business

Research, 64, 1195–1200.

Leech, N. L, Barrett, K. C, & Morgan, Jr. G. A. (2005). SPSS for Intermediate

Statistics: Use and Interpretation. New Jersey: Lawrence Erlbaum Associates,

Inc.

Leek, S., & Chansawatkit, S. (2006). Consumer confusion in the Thai mobile phone

market. Journal of Consumer Behaviour, 5, 518–532.

Levitt, T. (1980). Marketing success through differentiation-of anything. Harvard

Business Review, (Jan–Feb), 83–91.

Levy, S. J. (1959). Symbols for sale. Harvard Business Review, 37 (July–August),

117–119.

Little, J. A. R. (1988). A test for Missing Completely At Random for multivariate data

with missing values. Journal of the American Statistical Association, 83 (404),

1198–1202.

Loken, B., Ross, I., & Hinkle, R. L. (1986). Consumer ‘confusion’ of origin and brand

similarity perceptions. Journal of Public Policy & Marketing, 5, 195–211.

Lomax, W. E. Sherski, E., & Todd, S. (1999). Assessing the risk of consumer

confusion: Some practical test results. Journal of Brand Management, 7 (2),

119–132.

Lowe, M. C., & Damankos, F. J. (1968). Psychological and sociological dimension of

anomie in a psychiatric population. The Journal of Social Psychology, 74, 65–

74.

Lurie, N. H. (2004). Decision making in information-rich environments: The role of

information structure. Journal of Consumer Research, 30 (March), 473–486.

Lutz, R. J., & Kakkar, P. (1975). The psychological situation as a determinant of

consumer behaviour. Advances in Consumer Research, 2, 439–454.

Machleit, K. A., Eroglou, A. S., & Powell Mantel, S. (2000). Perceived retail crowding

and shopping satisfaction: What modifies this relationship? Journal of

Consumer Psychology, 9 (1), 29–42.

Mack, N., Woodsong, C., MacQueen, K. M., Guest, G., & Namey, E. (2005).

Qualitative research methods: A data collector’s filed guide. North Carolina:

Family Health International.

REFERENCES

347

Macklin, N. C. (1986). Classical conditioning effects in product/character pairings

presented to children, in Lutz, R. J. (Eds). Advances in Consumer Research, 13,

Association for Consumer Research, 198–203.

Maclnnis, J. D., & Jaworski, J. B. (1989). Information processing from advertisements.

Towards an integrative framework. Journal of Marketing, 53 (October), 1–23.

Malhotra, N. K. (1982). Information load and consumer decision making. Journal of

Consumer Research, 8 (4), 419–430.

Malhotra, N. K. (1984a). Information and sensory overload. Psychology & Marketing,

1 (3/4), 9–21.

Malhotra, N. K. (1984b). Reflections on the information overload paradigm in

consumer decision-making. Journal of Consumer Research, 10 (4), 436–440.

Malhotra, N. K., Arun, K. J., & Lagakos, S. W. (1982). The information overload

controversy: An alternative viewpoint. Journal of Marketing, 46 (Spring), 27–

37.

Malhotra, N. K., & Birks, D. F. (2007). Marketing research: An applied approach (3rd

Edition). London: Prentice Hall.

Malle, F. B., & Knobe, J. (1997). The folk concept of intentionality. Journal of

Experimental Social Psychology, 33, 101–121.

Markin, R. J., & Narayana, C. L. (1975). Behavior control: Are consumers beyond

freedom and dignity? Advances in Consumer Research, 3, 222–228.

Marshall, D., Anderson, A., Lean, M., & Foster, A. (1994). Healthy eating: Fruit and

vegetables in Scotland. British Food Journal, 96 (7), 18–24.

Mattila A. S., & Wirtz, J. (2001). Congruency of scent and music as a driver of in-store

evaluations and behaviour. Journal of Retailing, 77, 273–289.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of

organizational trust. Academy of Management Review, 20 (3), 709–734.

Mazaheri, E., Richard, M. O., & Laroche, M. (2011). Online consumer behaviour:

Comparing Canadian and Chinese website visitors. Journal of Business

Research, 64, 958–965.

Mazur, J.E. (2006). Learning and Behaviour (6th Edition). New Jersey: Pearson

Prentice Hall.

MacCallum, R. C., Widaman, K, F., Zhang, S., & Hong, S. (1999). Sample size in

factor analysis. Psychological Methods, 4 (1), 84–99.

Madden, C. S., Easley, R. W., & Dunn, M. G. (1995). How journal editors view

replication research. Journal of Advertising, 24 (Winter), 78–87.

REFERENCES

348

McAlister, L. (1982). A dynamic attribute satiation model of variety-seeking

behaviour. Journal of Consumer Research, 9 (September), 141–150.

McCarthy, E. J. (1964). Basic Marketing: A Managerial Approach (2nd Edition).

Homewood, Ill.: Richard D. Irwin.

McClosky, H., & Schaar, J. H. (1965). Psychological dimensions of anomy. American

Sociological Review, 30 (1), 14-40.

McColl-Kennedy, J. R., Patterson, P. G., Smith, A. K., & Brandy, M. K. (2009).

Customer rage episodes: Emotions, expressions and behaviors. Journal of

Retailing, 85 (2), 222–237.

McKenna, R. (1991). Marketing is everything. Harvard Business Review, 69 (1), 65–

79.

McNeill, P., & Chapman, S. (2005). Research Methods (3rd

Edition). Oxford:

Routledge.

Mehrabian, A., & Russell, J. A. (1974). An approach to environmental psychology.

Cambridge, MA: MIT Press.

Mendleson, N., & Polonsky, M. J. (1995). Using strategic alliances to develop credible

green marketing, Journal of Consumer Marketing, 12 (2), 4–18.

Merton, R. K. (1938). Social structure and anomie. American Sociological Review, 3

(June), 672–682.

Miaoulis, G., & D'Amato, N. (1978). Consumer confusion & trademark infringement.

Journal of Marketing, 42 (2), 48–55.

Micceri, T. (1989). The unicorn, the normal curve, and other improbable creatures.

Psychological Bulletin, 105 (1), 156–166.

Miceli, G., & Pieters, R. (2010). Looking more or less alike: Determinants of perceived

visual similarity between copycat and leading brands. Journal of Business

Research, 63 (11), 1121–1128.

Mick, D., G., & Fournier, S. (1998). Paradoxes of technology: Consumer cognizance,

emotions and coping strategies. Journal of Consumer Research, 25, 123–143.

Mick, D. G., Boniarczyk, S., & Haidt, J. (2004). Choose, choose, choose, choose,

choose, choose, choose: emerging and prospective research on the deleterious

effects of living in consumer hyperchoice. Journal of Business Ethics, 52 (2),

207–211.

Michael, J. (1980). On terms: The discriminative stimulus or SD. The Behavior

Analyst, 3, 47–49.

REFERENCES

349

Michael, J. (1986). Repertoire-altering effects of remote contingencies. The Analysis of

Verbal Behavior, 4, 10–18.

Mick, D. G. (2007). Degrees of freedom of will. An essential endless question in

consumer behavior. Journal of Consumer Psychology, 18, 17–21.

Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on

our capacity for processing information. The Psychological Review, 63 (2), 81–

97.

Mingers, J. (2001). Combining IS research methods: Towards a pluralist methodology.

Information System Research, 12 (3), 240–259.

Mingers, J., & Brocklesby, J. (1997). Multimethodology: Towards a framework for

mixing methodologies. Omega, 25 (5), 489–509.

Mintel (2009a). Ethical and green retailing, UK, 2009. London: Mintel. Available

online. URL:

http://academic.mintel.com/sinatra/oxygen_academic/search_results/show&/dis

play/id=280622 [Accessed 22 April, 2011].

Mintel (2009b). Online Grocery Retailing, UK, September 2009. London: Mintel.

Available online (using university passwords): URL

http://academic.mintel.com/sinatra/oxygen_academic/search_results/show&/dis

play/id=418439/display/id=480348#hit1 [Accessed 21 April 2011].

Mintel (2009c). PC and Laptops, UK, August, 2009. London: Mintel. Available online

(using university passwords): URL:

http://academic.mintel.com.abc.cardiff.ac.uk/display/474522/?highlight=true#

[Accessed 21 May 2012].

Mintel (2010). Food Retailing, UK, November 2010. London: Mintel. Available

online:

URL:http://academic.mintel.com/sinatra/oxygen_academic/display/id=479963/

display/id=554901 [Accessed 15 April, 2011].

Mintel (2013). Researching and buying technology products, UK, June 2013. London:

Mintel. Available online: URL

http://academic.mintel.com/sinatra/oxygen_academic/display/id=479963/displa

y/id=554901 [Accessed 10 June, 2013].

Mitchell, V.-W. (1998). A role for consumer risk perceptions in grocery retailing.

British Food Journal, 100 (4), 171–183.

Mitchell, V.-W. (1999). Consumer perceived risk: conceptualisation and models.

European Journal of Marketing, 33 (1/2), 163–195.

Mitchell, V.-W., & Kearney, I. (2002). A critique of legal measures of brand

confusion. Journal of Product & Brand Management, 11 (6/7), 357–379.

http://academic.mintel.com/sinatra/oxygen_academic/search_results/show&/display/id=280622

http://academic.mintel.com/sinatra/oxygen_academic/search_results/show&/display/id=280622

http://academic.mintel.com/sinatra/oxygen_academic/search_results/show&/display/id=418439/display/id=480348#hit1

http://academic.mintel.com/sinatra/oxygen_academic/search_results/show&/display/id=418439/display/id=480348#hit1

http://academic.mintel.com/sinatra/oxygen_academic/display/id=479963/display/id=554901




REFERENCES

350

Mitchell, V.-W., & Papavassiliou, V. (1997). Exploring consumer confusion in the

watch market. Market Intelligence & Planning, 15 (4), 164–171.

Mitchell, V.-W., & Papavassiliou, V. (1999). Marketing causes and implications of

consumer confusion. Journal of Product & Brand Management, 8 (4), 319–339.

Mitchell, V.-W., Walsh, G., & Yamin, M. (2005). Towards a conceptual model of

consumer confusion. Advances in Consumer Research, 32 (1), 143–150.

Moliner, P. & Tafani, E. (1997). Attitudes and social representations: a theoretical and

experimental approach. European Journal of Social Psychology, 27, 687–702.

Morgan, M. R., & Hunt, D. S. (1994). The commitment-trust theory of relationship

marketing. Journal of Marketing, 58 (July), 20–38.

Morgan, R. L., & Heise, D. (1988). Structure of emotions. Social Psychology

Quarterly, 51 (1), 19–31.

Murphy, E., & Dingwall, R. (2003). Qualitative Methods in Health Policy Research.

New York: Aldine de Gruyter.

Myers, J. H. (1996). Segmentation and positioning for strategic marketing decisions.

Chicago, IL: American Marketing Association.

Neelon, J. V., Champagne, T. M., Carlson, R. J., & Funk, G. S., (1996). The

NEECHAM confusion scale: construction, validation, and clinical testing.

Nursing Research, 45 (6), 324–330.

Neisser, U. (1967). Cognitive psychology. New York: Appleton-Century-Crofts.

Neisser, U. (1976). Cognition and reality: principles and implications of cognitive

psychology. San Francisco: WH Freeman and Company.

Nielsen Reports (2011). Online grocery shopping. London: Nielsen.

Nord, W. R., & Peter, P. J. (1980). A behavior modification perspective on marketing.

Journal of Marketing, 44, 36–47.

Nosek, A. B., Banaji, R. M., & Greenwald, G. A. (2002). E-research: Ethics, security

design and control in psychological research on the internet. Journal of Social

Issues, 58 (1), 161–176.

Office for National Statistics, (2012). Internet access - Households and individuals,

2012. Office for National Statistics. Available online, URL: <

http://www.ons.gov.uk/ons/dcp171778_275775.pdf>[Accessed February 2013].

Okouchi, H. (1999). Instructions as discriminative stimuli. Journal of the Experimental

Analysis of Behavior, 72 (2), 205–214.

http://www.ons.gov.uk/ons/dcp171778_275775.pdf

REFERENCES

351

Oliveira-Castro, J. M. (2013). Comments on Foxall’s ‘Intentionality, symbol and

situation in the interpretation of consumer choice’. Marketing Theory, 13 (1),

129–132.

Oppewal, H., & Koelemeijer, K. (2005). More choice is better: Effects of assortment

size and composition on assortment evaluation. International Journal of

Research in Marketing, 22, 45–60.

Ortony, A., Clore, G. L., & Foss, M. A. (1987). The referential structure of the

affective lexicon. Cognitive Science, 11, 341–364.

Ortony, A., Clore, G. L., & Collins, A. (1988). The Cognitive Structure of Emotions.

Cambridge: Cambridge Press.

Osgood, C.E. (1966). Dimensionality of the semantic space for communication via

facial expressions. Scandinavian Journal of Psychology, 7, 1–30.

O'Shaughnessy, J., & O'Shaughnessy, J., N. (2003). The Marketing Power of Emotion.

New York: Oxford University Press.

Pachauri, M. (2002). Consumer behaviour: a literature review. The Marketing Review,

2, 319–355.

Pal, J., & Byrom, J. (2005). More than one bite of the cherry: A study of UK grocery

retailers’ attempts to influence the retail planning policy making process.

Journal of Public Affairs, 5, 136–149.

Pallant, J. (2010). SPSS survival manual (4th

Edition). England: Open University Press.

Parasuraman, A., & Colby, C. L. (2001). Techno-ready marketing: How and why your

customers adopt technology. New York: The Free Press.

Parasuraman, A., Grewal, D., & Krishnan, R. (2004). Marketing Research. USA:

Houghton Mifflin Company.

Patton, M. Q. (2002). Qualitative research and evaluation methods. California, Sage

Publications, Inc.

Peter, J. P. (1979). Reliability: A review of psychometric basics and recent marketing

practices. Journal of Marketing Research, 16 (1), 6–17.

Peter, J. P., & Olson, J. C. (1983). Is science marketing? Journal of Marketing, 47

(Fall), 111–125.

Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion.

Advances in Experimental Social Psychology, Orlando, Florida: Academic

Press Inc.

REFERENCES

352

Petty, R. E., Cacioppo, J. T., & Schumann, D. (1983). Central and peripheral routes to

advertising effectiveness: the moderating role of involvement. Journal of

Consumer Research, 10, 135–146.

Peláez, M., & Moreno, R. (1999). Four dimensions of rules and their correspondence to

rule-governed behaviour: A taxonomy. Behavioural Development, 8 (1),

Spring, 21–27.

Phelps, A. E. (2006). Emotion and cognition: insights from studies of the human

amygdala. Annual Review of Psychology, 57, 27–53.

Plutchik, R. (1980). Emotion: a psychoevolutionary synthesis. New York: Harper and

Row.

Plutchik, R. (1985). On emotion: The chicken-and-egg problem revisited. Motivation

& Emotion, 9 (2), 197–200.

Plutchik, R. (1994). The Psychology and Biology of Emotion. New York: Harper

Collins.

Poiesz, T. B. C. (2004). The free market illusion psychological: limitations of

consumer choice. Journal of Economic Psychology, 49 (2), 309–338.

Poppen, R.L. (1989). Some clinical implications of rule-governed behaviour. In S. C.

Hayes. (Eds). Rule governed behaviour: Cognition, contingencies, and

instructional control. Reno, NV: Context Press. pp. 325–357.

Potter, J. & Wetherell, M. (1987). Discourse and Social Psychology: beyond attitudes

and behaviour. London: Sage.

Rafiq, M., & Collins, R. (1996). Lookalikes and customer confusion in the grocery

sector: An exploratory survey. International Review of Retail, Distribution &


Rajala, A. K., & Hantula, D. A. (2000). Towards a behavioral ecology of consumption:

Delay-reduction effects on foraging in a simulated Internet mall. Managerial

and Decision Economics, 21, 145–158.

Raju, P. S. (1980). Optimum stimulation level: Its relationship to personality

demographics and exploratory behaviour. Journal of Consumer Research, 7


Ratner, R. K., & Kahn, B. E. (2002). The impact of private vs. public consumption on

variety seeking behaviour. Journal of Consumer Research, 29 (September),

246–258.

Richins, M. L. (1997). Measuring emotions in the consumption experience. Journal of


REFERENCES

353

Ritov, I., & Baron, J. (1992). Status-quo and omission biases. Journal of Risk and

Uncertainty, 5, 49–61.

Ritzer, G. (2004). The McDonaldization of society. Revised New Century Edition

Thousand Oaks, CA: Sage Publishing.

Rolls, E. T. (2000). Précis of the brain and emotion. Behavioral and Brain Sciences,

23, 177–234.

Rolls, E. T. (2005). Emotion explained. Oxford, UK: Oxford University Press.

Roseman, I. J. (1991). Appraisal determinants of discrete emotions. Cognition and

Emotion, 5, 161–200.

Roseman, J. I., Antoniou, A. A., & Jose, J. P. (1996). Appraisal determinants of

emotions: constructing a more accurate and comprehensive theory. Cognition &

Emotion, 10 (3), 241–277.

Roseman, I. J., & Smith, C. A. (2001). Appraisal Theory: Overview, assumptions,

varieties, controversies. In: K. Scherer, A. Schorr, T. Johnstone (Eds.).

Appraisal Processes in Emotion: Theory, Methods, Research. Oxford: Oxford

University Press.

Rotter, J. B. (1966). Generalized expectancies for internal versus external locus of

control of reinforcement. Psychological Monographs, 80, 1–28.

Rozin, P., & Cohen, B. A. (2003a). High frequency of facial expression corresponding

to confusion, concentration and worry in an analysis of naturally occurring

facial expressions of Americans. Emotion, 3 (1), 68–75.

Rozin, P., & Cohen, B. A. (2003b). Confusion infusions, suggestives, correctives, and

other medicines. Emotion, 3 (1), 92–96.

Rudestam, K. E., & Newton, R. R. (2007). Surviving your dissertation: a

comprehensive guide to content and process. (3rd

Edition). California: Sage

Publications Inc.

Russell, J. A. (1978). Evidence of convergent validity on the dimensions of affect.

Journal of Personality and Social Psychology, 36 (10), 1152–1168.

Russell, J. A. (1979). Affective space is bipolar. Journal of Personality and Social

Psychology, 37 (3), 345–356.

Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social

Psychology, 39 (6), 1161–1178.

Russell, J. A., & Mehrabian, A. (1977). Evidence for a three-factor theory of emotions.

Journal of Research in Personality, 11, 273–294.

REFERENCES

354

Russell, J. A., & Mehrabian, A. (1978). Approach-avoidance and affiliation as

functions of the emotion-eliciting quality of an environment. Environment and

Behavior, 10 (3), 355–387.

Russell, J. A., & Mehrabian, A. (1997). Evidence for a three-factor theory of emotions.

Journal of Research in Personality, 11 (3), 273–294.

Russell, J. A., & Pratt, G. (1980). A description of the affective quality attributed to

environments. Journal of Personality and Social Psychology, 38 (2), 311–322.

Russo, J. (1974). More information is better: A Re-evaluation of Jacoby, Speller and

Kohn. Journal of Consumer Research, 1 (December), 68–72.

Ruth, J. A, Brunel, F. F, & Otnes, C. C. (2002) Linking thoughts to feelings:

investigating cognitive appraisals and consumption emotions in a mixed-

emotions context. Journal of the Academy of Marketing Science, 30 (1), 44–58.

Ryan, R. M., & Deci, L. D. (2001). On happiness and human potentials: A review of

research on hedonic and eudaimonic well-being. Annual Review of Psychology,

52, 141–166.

Saunders, M., Lewis, P., & Thornhill, A. (2009). Research methods for business

students (5th

Edition). New York: Prentice Hall.

Scammon, D. L. (1977). Information load and consumers. Journal of Consumer

Research, 4 (December), 148–155.

Scherer, R. K. (2005). What are emotions? And how can they be measured? Social

Science Information, 44 (4), 695–729.

Schlinger, H. D. (1993). Separating discriminative and function-altering effects of

verbal stimuli. The Behavior Analyst, 16, 9–23.

Schlinger, H. D. (2008). Conditioning the behaviour of the listener. International

Journal of Psychology and Psychological Therapy, 8 (3), 309–322.

Schlinger, H. D., & Blakely, E. (1987). Function-altering effects of contingency

specifying stimuli. The Behavior Analyst, 10 (1), 41–45.

Scheibehenne, B., Greifender, R., & Todd, P. M. (2010). Can there be too many

options? A meta-analytic review of choice overload. Journal of Consumer

Research, 37, 409–425.

Schmitt, B. (1999). Experiential marketing. Journal of Marketing Management, 15,

53–67.

Schwartz, B. (2000). Self-determination: The tyranny of freedom. American

Psychologist, 55, 79–88.

Schwartz, B. (2004). The paradox of choice. New York: Ecco.

REFERENCES

355

Schwartz, B., Ward, A., Monterosso, J., Lyubomirsky, S., White, K., & Lehman, D. R.

(2002). Maximizing versus satisficing: Happiness is a matter of choice. Journal

of Personality and Social Psychology, 83 (5), 1178–1197.

Schweizer, M. (2004). Consumer confusion im handel–Ein umweltpsychologisches

erklärungsmodell. St. Gallen: Rosch-Buch, Stresslitz.

Schweizer, M., & Rudolph, T. (2004). Wenn käufer streiken–Mit klarem profil gegen

consumer confusion und kaufmüdigkeit. Wiesbaden: Gabler.

Schweizer, M., Kotouc, A. J., & Wagner, T. (2006). Scale development for consumer

confusion. Advances in Consumer Research, 33 (1), 184–190.

Searle, J. R. (1983). Intentionality. Cambridge: Cambridge University Press.

Searle, J. R. (1990). Collective intentions and actions. In P. Cohen, J. Morgan and M.

Pollack (Eds). Intentions in Communication. Cambridge, MA: A Bradford

Book. The MIT Press. pp. 401–415.

Searle, J. R. (2010). Making the social world: The structure of human civilization. New

York: Oxford University Press Inc.

Settle, R. B., & Alreck, L. P. (1988). Hyperchoice shapes the marketplace. Marketing

Communications, 13(May), 61, 15–20.

Shankar, A., & Patterson, M. (2001). Interpreting the past, writing the future. Journal

of Marketing Management, 17, 481–501.

Sigurdsson, V., Foxall, G. R., & Saevarsson, H. (2010). In-store experimental approach

to pricing and consumer behavior. Journal of Organizational Behavior

Management, 30, 234–246.

Simon, H. A. (1955). A behavioral model of rational choice. Quarterly Journal of

Economics, 69 (1), 99–118.

Singer, E., Hippler, J. H., & Schwarz, N. (1992). Confidentiality assurances in surveys:

reassurance or threat? International Journal of Public Opinion Research, 4 (3),

256–268.

Singer, E., von Thurn, D. R., & Miller, E. R. (1995). Confidentiality assurances and

response: A quantitative review of the experimental literature. Public Opinion

Quarterly, 59, 66–77.

Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. (1987). Emotion knowledge:

further exploration of a prototype approach. Journal of Personality and Social

Psychology, 52, 1061–1086.

Slater, R., & Lipman, A. (1977). Staff assessment of confusion and the situation of

confused residents in homes for old people. The Gerontologist, 17 (6), 523–530.

REFERENCES

356

Skinner, B. F. (1950). Are theories of learning necessary? Psychological Review, 57,

193–216.

Skinner, B. F. (1957). Verbal behavior. Englewood Cliffs, NJ: Prentice-Hall.

Skinner, B. F. (1966). An operant analysis of problem-solving. In B. Kleinmuntz (Ed.),

Problem solving: Research, method, teaching (pp. 225–257). New York: Wiley.

Skinner, B. F. (1969). Contingencies of Reinforcement: A Theoretical Analysis.

Enlewood Cliffs. New Jersey: Prentice-Hall.

Skinner, B. F. (1974). About Behaviorism. New York: Alfred A. Knopf.

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive appraisal in emotion.

Journal of Personality and Social Psychology, 48, 813–838.

Smith, C. A., & Lazarus, S. R. (1992). Emotion and Adaptation, in L. A., Pervin (ed.).

Handbook of Personality: Theory and Research, (pp. 609–637), New York:

Guildford.

Smith, A. K, & Bolton, R. N. (2002). The effects of customers’ emotional responses to

service failures on their recovery effort evaluations and satisfaction judgements.

Journal of the Academy of Marketing Science, 30 (1), 5–23.

Solomon, R. M., Barmossy, G., Askegaard, S., Hogg, K., M. (2010). Consumer

behaviour. A European perspective (4th

Edition). London: Prentice Hall Inc.

Soriano, M. Y., Foxall, G. R. & Pearson, G. J. (2002). Emotion and environment: a test

of the Behavioural Perspective Model in a Latin American context. Journal of

Consumer Behaviour, 2 (2), 138–154.

Soriano, M. Y., & Foxall, G. R. (2006). The emotional power of place: The fall and

rise of dominance in retail research. Journal of Retailing and Consumer

Services, 13, 403–416.

Soriano, M. Y., Foxall, G. R., & Newman, J. A. (2013). The impact of the interaction

of utilitarian and informational reinforcement and behaviour setting scope on

consumer response. Psychology and Marketing, 30 (2), 148–159.

Spangenberg, E. R., Grohmann B., &, Sprott D, E. (2005). It's beginning to smell (and

sound) a lot like Christmas: the interactive effects of ambient scent and music in

a retail setting. Journal of Business Research, 58 (11), 1583–1589.

Spiggle, S. (1994). Analysis and interpretation of qualitative data in consumer research.

Journal of Consumer Research, 21 (December) 491–503.

Sproles, G. B. & Kendall, E. (1986). A methodology for profiling consumers' decision-

making styles. Journal of Consumer Affairs, 20 (2), 267–279.

REFERENCES

357

Stebbins, A., R. (2001). Exploratory Research in the Social Sciences. London: Sage

Publications Inc.

Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of

the self. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology,

21, 261–302. San Diego, CA: Academic Press.

Steenkamp, J-B.E.M., & van Trijp, H.C.M. (1991). The use of LISREL in validating

marketing constructs. International Journal of Research in Marketing, 8, 283–

299.

Steenkamp, J. B., & Baumgartner, H. (1992). The role of optimum stimulation level in

exploratory consumer behaviour. Journal of Consumer Research, 19


Sternberg, R. (1999). The nature of cognition. Cambridge, MA: MIT Press.

Stokols, D. (1995). The paradox of environmental psychology. American Psychologist,

50 (10), 821–837.

Storm, C., & Storm, T. (1987). A taxonomic study of the vocabulary of emotions.

Journal of Personality and Social Psychology, 53, 805–816.

Summers, J. O. (1974). Less information is better? Journal of Marketing Research, 11

(November), 467–468.

Szmigin, I., & Foxall, G. (2000). Interpretive consumer research: how far have we

come? Qualitative Market Research, 3 (4), 187–189.

Tabachnick, G. B., & Fidell, L. S. (2007). Using Multivariate Statistics (5th

Edition).

Boston, USA: Pearson Education Inc.

Taylor, L. J. (1968). Alienation, anomie and delinquency. British Journal of Social and

Clinical Psychology, 7 (2), 93–105.

Taylor, S. E., & Brown, J. D. (1988). Illusion and wellbeing: A social psychological

perspective on mental health. Psychological Bulletin, 103, 193–210.

The British Psychological Society, (2009). Code of ethics and conduct, online: <URL:

http://www.bps.org.uk/sites/default/files/documents/code_of_ethics_and_condu

ct.pdf > [Accessed July 2012].

Titus, P. A., & Everett, P. B. (1996). Consumer way finding tasks, strategies, and

errors: An exploratory field study. Psychology & Marketing, 13 (3), 265–290.

Thompson, C. J., & Hirschman, E. C. (1995). Understanding the socialised body: a

poststructuralist analysis of consumers' self-conceptions, body images, and self-

care practices. Journal of Consumer Research, 25 (2), 139–153.

http://www.bps.org.uk/sites/default/files/documents/code_of_ethics_and_conduct.pdf

http://www.bps.org.uk/sites/default/files/documents/code_of_ethics_and_conduct.pdf

REFERENCES

358

Thompson, C. J., Locander, W. B., & Pollio, H. R. (1990). The lived meaning of free

choice: an existential phenomenological description of everyday consumer

experiences of contemporary married women. Journal of Consumer Research,

17 (December), 346–361.

Törneke, N., Luciano, C., & Salas, S. V. (2008). Rule-governed behaviour and

psychological problems. International Journal of Psychology and

Psychological Therapy, 8 (2), 141–156.

Turley, L. W., & Milliman, R. E. (2000). Atmospheric effects on shopping behavior: a

review of the experimental evidence. Journal of Business Research, 49 (2),

193–211.

Turnbull, P. W., Leek, S., & Ying, G. (2000). Customer confusion: The mobile phone

market. Journal of Marketing Management, 16 (1/3), 143–163.

Tversky, A., & Shafir, E. (1992). Choice under conflict: The dynamics of deferred

decision. Psychological Science, 3 (6), 358–361.

Urbany, J.E., Dickson, P.R., & Wilkie, W. L. (1989). Buyer uncertainty and

information search. Journal of Consumer Research, 16 (Sept), 208–215.

van Birgelen, M., de Ruyter, K., & Wetzels, M. (2001). Customer evaluations of after-

sales service contact modes: an empirical analysis of national culture’s

consequences. International Journal of Research in Marketing, 19 (1), 43–64.

Vaughan, E. M. (1985). Repeated acquisition in the analysis of rule-governed

behaviour. Journal of the Experimental Analysis of Behaviour, 44 (2), 175–184.

Venkatesh, A. (1992). Postmodernism, consumer culture and the society of the

spectacle. Advances in Consumer Research, 19, 199–202.

Verbeke, W., & Bagozzi, R. P. (2003). Exploring the role of self- and customer-

provoked embarrassment in personal selling. International Journal of Research

in Marketing, 20, 233–258.

Walsh, G., & Mitchell, V.-W. (2005a). Consumer vulnerability to perceived product

similarity problems: Scale development and identification. Journal of

Macromarketing, 25 (2), 140–152.

Walsh, G., & Mitchell, V.-W. (2005b). Demographic characteristics of consumers who

find it difficult to decide. Marketing Intelligence and Planning, 23 (3), 281–

295.

Walsh, G., & Mitchell, V.-W. (2010). The effect of consumer confusion proneness on

word of mouth, trust, and consumer satisfaction. European Journal of

Marketing, 44 (6), 838–859.

REFERENCES

359

Walsh, G., Hennig-Thurau, T., & Mitchell, V.-W. (2007). Consumer confusion

proneness: Scale development, validation, and application. Journal of

Marketing Management, 23 (7/8), 697–721.

Walsh, G., Mitchell V.-W., & Frenzel, T. (2004). Consumer e-confusion on the

Internet. Thexis, June, 17–21.

Walsh, G., Shiu, E., Hassan, L.M., Michaelidou, N., & Beatty, S. E. (2011). Emotions,

store-environmental cues, store choice criteria, and marketing outcomes.

Journal of Business Research, 64, 737–744.

Ward, C. J., & Barnes, W. J. (2001). Control and affect: the influence of feeling in

control of the retail environment on affect, involvement, attitude and behaviour.

Journal of Business Research, 54, 139–144.

Warlop, L., & Alba, J. W. (2004). Sincere flattery: trade-dress imitation and consumer

choice. Journal of Consumer Psychology, 14 (1/2), 21–27.

Watson, J. (1913). Psychology as the behaviourist sees it. Psychological Review, 20,

158–177.

Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood.

Psychological Bulletin, 98, 219–235.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief

measures of positive and negative affect: the PANAS scales. Journal of

Personality and Social Psychology, 54, 1063–1070.

Wells, W. D. (2001). The perils of N=1. Journal of Consumer Research, 28(3), 494–

498.

West, E. G., Bruno, L., Carole, G., & Shannon, L. S. (2002). Consumer confusion over

the significance of meat attributes: The case of veal. Journal of Consumer

Policy, 25, 65–88.

Westbrook, R. A. (1987). Product/ consumption-based affective responses and post-

purchase processes. Journal of Marketing Research, 24, 258–270

Wilk, R. R. (2001). The impossibility and necessity of re-inquiry: finding middle

ground in social research. Journal of Consumer Research, 28 (2), 308–312.

Wilkie, W. L. (1974). Analysis of effects of information load. Journal of Marketing

Research, 11 (November), 462–466.

Wiseman, J. (1994). Beliefs about food components, foods, fat and heart disease in

New Zealand. British Food Journal, 96 (11), 14–19.

Woodward, S. E., & Hall, E. R. (2010). Consumer confusion in the mortgage market:

Evidence of less than a perfectly transparent and competitive market. American

Economic Review, 100 (2), 511–515.

REFERENCES

360

Wright, P. (1975). Consumer choice strategies—Simplifying vs. Optimizing. Journal

of Marketing Research, 12 (February), 60–67.

Yi, Y. (1990). A critical review of consumer satisfaction. In V. A. Zeithaml (Ed.)

Review of Marketing. Chicago: American Marketing Association, 68–123.

Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and

Social Psychology, 9 (2), 1–27 [Monograph].

Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American

Psychologist, 35, 151–175.

Zajonc, R. B. (1984). On the primacy of affect. American Psychologist, 39, 117–123.

Zelenski, M. J., & Larsen, J. R. (2000). The distribution of basic emotions in everyday

life: A state and trait perspective from experience sampling data. Journal of

Research in Personality, 34 (2), 178–197.

Zettle, R. D. (1990). Rule-governed behavior: A radical behavioral answer to the

cognitive challenge. The Psychological Record, 40, 41–49.

Zettle, R. D., & Hayes S. C. (1982). Rule-governed behaviour: A potential theoretical

framework for cognitive behavioural therapy. Advances in Cognitive-

Behavioural Research and Therapy, 1, 73–117.

Zimet, G., D., Dahlem, N., W., Zimet, S., G., & Farley G., K., (1988). The

multidimensional scale of perceived social support. Journal of Personality

Assessment, 52 (1), 30–41.

Zuckerman, M. (1979). Sensation seeking. Hillsdale, NJ: Lawrence Erlbaum

Association.

Zuckerman, M., Porac, J., Lathin, D., Smith, R., & Deci, E. L. (1978). On the

importance of self-determination for intrinsically motivated behaviour.

Personality and Social Psychology Bulletin, 4, 443–446.

APPENDIX

361

12. APPENDIX

APPENDIX

362

Appendix 1

Table 1.1: Selected studies related to consumer confusion (presented in chronological order based on the year of publication).

Key areas of study

Conceptual and

empirical papers

Type of

paper Core Aim

Conceptualisation of

consumer confusion

Quasi-

Conceptualisation

Area/

Focus

Confusion

triggers

Consumer

characteristic

s

Confusion

consequences

Jacoby, Speller,

& Kohn, 1974a Empirical

Examine the

information

overload

implications for

brand choice in a

consumer

context.

‘…feeling of

confusion, of not

having obtained

the best buy, and

feeling that another

brand was better.’

(p. 66)

Overload

confusion +

Miaoulis & D’

Amato, 1978 Empirical

Examine the

research issue of

consumer

confusion

tailored to a

specific type of

trademark

infringement

problem (the

Tic-Tac case).

‘We take the position

here that confusion is in

effect stimulus

generalization.’(p. 49)

Brand

confusion +

Loken et al.,

1986 Empirical

To test whether

similarity in

physical

appearance

between two

brands

influences

consumer

perception of a

common

business origin.

‘(…) physical

similarities between

products may result in

the misattribution of

source of origin or

identity by the

consumer.’ (p. 196)

Brand

confusion +

APPENDIX

363

Sproles &

Kendall, 1986 Empirical

Conceptualise

eight basic

consumer

decision-making

characteristics

and develop a

Consumer

Inventory Style

to measure them

empirically.

Confused by

overchoice

consumer is one

of these styles.

‘Consumers

perceive many

brands and stores

from which to

choose and have

difficulty making

choices.

Furthermore, they

experience

information

overload.’ (p. 274)

Consumer

confusion +

Poiesz &

Verhallen, 1989

Similarity in

advertisements

and commercial

messages.

‘Brand confusion is a

phenomenon that occurs

at the individual level

(…) and is

predominantly non-

conscious in nature.’ (p.

233)

Brand

confusion (in

advertising)

Foxman et al.,

1990 Empirical

Examines

individual

characteristics as

an indication of

predisposition to

confusion and

why some

consumers

appear more

likely to become

confused than

others.

‘(…) consumers

who are mislead

clearly are

confused’(p. 172)

Brand

confusion + +

APPENDIX

364

Foxman et al.,

1992 Conceptual

A definition and

theoretical

framework for

consumer brand

confusion are

developed,

broadening the

concept of brand

confusion.

‘(…) consists of one or

two errors in inferential

processing that lead a

consumer to

unknowingly form

inaccurate beliefs about

the attributes or

performance of a less

known brand based on a

more familiar brand’s

attributes or

performance.’ (p. 125)

Brand

confusion + +

Kapferer, 1995 Empirical

Test for the

presence of

confusion

caused be

lookalikes by

means of a

tachistoscopic

experiment.

‘(…) close imitation of a

national brand (…) on

the basis of which

consumers may make

inferences and

attributions of similarity

of use, of content, if not

of origin.’ (p.551)

Brand

(imitation)

confusion +

Rafiq & Collins,

1996 Empirical

Exploratory

study

investigating the

effects of look-

alike own brand

products on the

possibility of

consumer

confusion in the

grocery sector.

‘Brand owners

have claimed that

look-alike own

label products

confuse

consumers’ (p.

329)

Brand

confusion

Balabanis &

Craven, 1997 Empirical

An exploratory

investigation of

consumer

confusion

caused by look-

alike own brand

products (p. 300)

‘(...) suggest that

shoppers under

different (...)

circumstances are

likely to become

confused by

lookalikes’.

Brand

confusion

APPENDIX

365

Kohli & Thakor,

1997 Conceptual

Provide a

framework,

helpful in the

brand name

creation process.

‘(…) confusion,

when consumers

may pick

confusingly similar

names, instead of

the target names.’

(p. 213)

Brand-name

confusion

Mitchell &

Papavassiliou,

1997

Empirical

Sources of

confusion in the

watch market

and empirically

describe how

buyers cope with

it.

No clear definition

provided.

Conceptualising

confusion through a

portrayal of its

antecedents (p. 165)

‘(…) confusion

from over choice

(…). These

features can

overload

consumers who are

often confused

about how to use

these features for

their benefits’.

Consumer

confusion + +

Jacoby &

Morrin, 1998 Conceptual

Review several

trademark

infringement

cases (1994–

1997) and make

observations

regarding trends

in this domain

and areas for

further research.

‘If someone other

than the owner

were to use a

trademark, there

would be the

possibility that

such use (by the

second or junior

user) could cause

consumers to be

confused regarding

who actually

makes the

product.’ (p.97)

+

Huffman &

Kahn, 1998 Empirical

Provide an

understanding of

how retailers

should expose

customers to

product

information

‘The huge number

of potential options

(…) may be

confusing’ and

‘The confusion a

consumer

experiences with a

Confusion

caused be

extended

assortments

+ +

APPENDIX

366

(attribute-based

or alternative-

based)

particularly in a

high variety

situation.

wide assortment of

options, however is

due to the

perceived

complexity, not

necessarily to the

actual complexity

or variance.’ (p.

492–493)

Mitchell &

Papavassiliou,

1999

Conceptual

Conceptual

paper which

initiated a

holistic

consideration of

consumer

confusion.

‘Confusion is more than

subconscious mistakes; it

is a state of mind which

affects information

processing and decision

making. The consumer

may therefore be aware

or unaware of

confusion.’ ( p. 327)

Consumer

confusion + + +

Chryssochoidis,

2000 Empirical

Examines the

influence of

consumer

confusion on

late introduced

differentiated

products using

data regarding

organic food

products.

‘(…) confusion is

defined as a situation in

which

consumers form

inaccurate beliefs about

the attributes or

performance of a less

known product as they

base themselves on a

more familiar product’s

attributes or

performance.’ (p. 705)

Product

confusion + + +

Turnbull et al.,

2000 Empirical

Identify the

extent of

consumer

confusion and

understand its

impact on both

the mobile

phone industry

and its

‘(…) consumer

confusion is defined as

consumer failure to

develop a correct

interpretation of various

facets of a

product/service, during

the information

processing procedure.’

Consumer

confusion + +

APPENDIX

367

consumers. (p. 145)

Mitchell &

Kearney, 2002 Conceptual

A critique of

legal measures

of consumer

confusion.

‘Consumer confusion

concerns a type of

subjective experience

(i.e. an unpleasant state

of mental discomfiture)

relating to an object,

usually a brand, that

affects the overall

evaluation of that object.

Thus, we argue that

consumer confusion is

an attitude and as such

can be viewed as having

behavioural, cognitive

and affective

components (…)’ (p.

357)

Brand

confusion +

West et al., 2002 Empirical

Consumer

confusion over

the significance

of meat (veal)

attributes.

Confusion is used

at the general level

without

specifically

defining it e.g.

‘confusion over the

significance of

extrinsic labelling

of production

method appeared

to be greater than

confusion over the

significance of

colour.’ (p. 83)

Consumer

confusion +

Liefeld, 2003 Empirical

To investigate

the validity of

survey methods

estimating the

likelihood of

confusion (in

cases of

Brand

similarity and

confusion

(trademark

infringement

cases)

+

APPENDIX

368

trademark

infringement)

Schweizer, 2004

Empirical

To introduce and

test a consumer

confusion

framework

based on the

principles of

environmental

psychology.

‘Consumer confusion is

an emotionally laden,

dysfunctional state of

mind, which makes it

difficult for consumers

to efficiently and

effectively select and

interpret stimuli.’ (p. 34)

Consumer

confusion + + +

Mitchell et al.,

2005

Conceptual

Form a

conceptual

model of

consumer

confusion based

on relevant

theory and

previous studies.

Proposed a conceptual

model of consumer

confusion

‘(…) conceptualizing

confusion as having

three consequences i.e.

cognitive, affective and

behavioural, which we

suggest are positively

correlated, irrespective

of the antecedents’

confusion experienced.’

(p. 143)

Consumer

confusion + + +

Walsh &

Mitchell, 2005 Empirical

Develop a

parsimonious

scale for

measuring

consumers’

orientation for

inferring that all

products within

a category are

similar and

identify

consumers

vulnerable to it.

‘(...) in this case is

the mistaken

purchase, misuse,

misunderstanding,

or misattribution of

various product

attributes caused

by thinking that

products are

similar when they

are different.’ (p.

140)

Product

(similarity)

confusion + +

APPENDIX

369

Schweizer et al.,

2006 Empirical

Develop a

relevant scale of

consumer

confusion based

on the principles

of environmental

psychology. The

aim has been to

investigate all

possible

confusion

triggers.

‘(...) we define the

phenomenon as a result

of a temporary

exceedance of an

individual capacity

threshold for absorbing

and processing

environment stimuli.

Consumer confusion is

an emotional state that

makes it difficult for

consumers to select and

interpret stimuli.’ (p.

185)

Consumer

confusion +

Walsh et al.,

2007

Empirical

Conceptualise

consumer

confusion

proneness,

provide 1. a new

scale to measure

it and 2.

empirical

evidence on how

it affects

consumer

behaviour.


proneness is a

multidimensional

phenomenon (similarity,

ambiguity and overload

confusion proneness)

that has a significant

impact on purchase

postponement and

loyalty behaviour.’ (p.

713)

Consumer

confusion + +

Walsh &

Mitchell, 2010 Empirical

Access the effect

of consumer

confusion

proneness (see

Walsh et al.,

2007) on word

of mouth, trust

and consumer

satisfaction.


proneness is a

multidimensional

phenomenon (similarity,

ambiguity and overload

confusion proneness)

that has a significant

impact on purchase

postponement and

loyalty behaviour.’

(based on Walsh et al.,

2007)

Consumer

confusion + +

APPENDIX

370

Kasper et al.,

2010 Empirical

The paper

investigates

whether

consumers who

face different

degrees of

confusion use

different coping

strategies

depending upon

their decision-

making styles.

‘Thus in our

conceptualisation,

consumer confusion is a

self-reported overload,

and as such a conscious

phenomenon.’ (p. 142)

Overload

confusion + +

Source: this study. Adapted and extended from Mitchell et al., 2005.

Appendix

371

Appendix 2

Table 2.1: Rule-Governed Behaviour- Units of Analysis Case I: Speaker and Listener as separate individuals

Units of speaker behaviour Description Example/ Notes

Tacts

A verbal operant in which a

response of given form is

evoked or strengthened by a

particular object or event or

their properties.

e.g. ‘That’s a chair’. A chair

remains a chair regardless of

my needs or desires.

A tact is not the same as

naming or referring to. A tact

requires direct experience, it is

experiential.

Mands

A verbal operant in which the

response is under the control

of conditions of reinforce-

ability in the speaker and of

antecedents indicating an

availability of a relevant

consequence.

e.g. A person enters a room and

wanting to sit down says: ‘a

chair’. A chair is then brought

forth by a listener. A mand is

then reinforced by

characteristic listener-mediated

consequences.

Note: Tacts often look like descriptions (that’s water) and mands often look like requests (bring

me water).

Units of listener behaviour Description Example/ Notes

Pliance

Rule-governed behaviour

primarily under the control of

apparent speaker-mediated

consequences for a

correspondence between the

rule and the relevant

behaviour.

The rule itself is called a Ply.

Drawn from the world

compliance.

e.g. A thief says: Your wallet or

your life. The listener hands the

thief the wallet. They imply an

expected action from the

listener and might contain

words like would and should.

Tracking


under the control of the

apparent correspondence

between the rule and the way

the world is arranged.

The rule itself is called a

Track. Drawn from following a

path.

e.g. An advice: The way to get

to Cardiff is through M4.

Augmenting


under the control of apparent

changes in the capacity of

The rule itself is termed

Augmental. It suggests a

changed or heightened state of

Appendix

372

events to function as

reinforcers or punishers.

affairs.

e.g. A child that follows the

rule ‘Eat vegetables and you

will be a strong boy’ because of

the future consequence of

becoming a strong boy follows

an augmental (Törneke et al.,

2008)

Case II: Self-rules

Self-pliance

It occurs when we act as if

the rule is to be followed

simply because it has been

formulated.

e.g. No matter what happens, I

am never going to speak to a

woman again- even if I know

she’s good for me.

Self-tracking

When we act as if the rule is

to be followed because it is a

description of the state of

affairs.

e.g. I think Ann dislikes me. I

will ignore her and that way, I

will not get hurt.

Self-augmenting

When the speaker attempts to

induce emotional changes in

themselves.

e.g. A woman who is

ambivalent about having an

abortion, might repeat to herself

poems and songs to be able to

reach an easier decision.

Source: Explanations, examples and notes are taken from Zettle & Hayes, 1982, p. 79-92. The

information is here summarised and presented in a table format.

Appendix

373

Appendix 3

Table 3.1. Comparison of quantitative and qualitative research approaches

Quantitative Qualitative

General framework Seek to examine

hypotheses about

phenomena.

Seek to explore

phenomena.

Instruments use more

rigid style of eliciting

and categorising

responses to questions.

Instruments use more

flexible, iterative style

of eliciting and

categorising responses

to questions.

Use structured methods

such as questionnaires,

surveys and structured

observations.

Use semi-structured

methods such as in-

depth interviews, focus

groups, and participant

observation.

Analytical objectives To quantify variation To describe variation.

To predict relationships. To describe and explain

relationships.

To describe

characteristics of the

population.

To describe individual

experiences.

To describe group

norms.

Question format Closed-ended Open-ended

Data format Numerical Textual

Flexibility in study design Study design is stable. Some aspects of the

study are flexible (e.g.

the wording of

questions depending on

the situation).

Participants’ responses

do not influence or

determine how and

which questions

researchers ask next.

Participants’ responses

influence how and

which responses to be

asked next.

Study design is subject

to statistical

assumptions and

conditions.

Study design is

iterative.

Source: Mack et al. 2005, p. 3

Appendix

374

Appendix 4

Table 4.1: Pilot Test 2- Sample demographic profile.

Sample (n=56)

n %

Gender Males 22 39.3

Females 34 60.7

Age 18–24 1 1.8

25–34 16 28.6

35–44 9 16.1

45–54 13 23.2

55–64 16 28.6

65+ 1 1.8

Higher completed

education

GCSE/ A-levels 4 7.1

Vocational or Technical School 1 1.8

Higher Education 16 28.6

Postgraduate Degree 35 62.5

Other - -

Ethnic Group White 46 82.1

Mixed 3 5.4

Asian (Asian British) 5 8.9

Black (Black British) 1 1.8

Other 1 1.8

Working Status Employed full-time (30+ hours per week) 42 75

Employed part time (less than 30hours per

week)

13 23.2

Unemployed - -

Student 1 1.8

Retired - -

Self-employed - -

Housewife/husband - -

Household Size 1 person 13 23.2

2 person 20 35.7

More than 2 persons 23 41.0

Channel for buying

groceries

In store 50 10.7

Online 6 89.3

Channel for buying

PC/laptops

In store 23 58.9

Online 33 41.07

Source: this pilot study

Appendix

375

Appendix 5

Table 5.1. Count and percentages of missing items

N Missing

Count Percent

P1 520 0 .0

P2 520 0 .0

D1 520 0 .0

P3 520 0 .0

A1 520 0 .0

P4 520 0 .0

P5 520 0 .0

P6 520 0 .0

A2 520 0 .0

D2 519 1 .2

A3 520 0 .0

A4 520 0 .0

D3 520 0 .0

A5 520 0 .0

D4 519 1 .2

D5 519 1 .2

A6 519 1 .2

D6 520 0 .0

AP1 520 0 .0

AV1 516 4 .8

AP2 517 3 .6

AV2 517 3 .6

AP3 518 2 .4

AV3 516 4 .8

S1 520 0 .0

S2 520 0 .0

S3 520 0 .0

O1 520 0 .0

S4 520 0 .0

O2 520 0 .0

O3 520 0 .0

O4 520 0 .0

AM1 520 0 .0

AM2 520 0 .0

AM3 520 0 .0

AM4 520 0 .0

O5 520 0 .0

MEx1 520 0 .0

Age 520 0 .0

Gender 518 2 .4

Education 514 6 1.2

Ethnic group 520 0 .0

Household 514 6 1.2

Working status 518 2 .4

Buying groceries 510 10 1.9

Buying pc/laptops 518 2 .4

Source: this study

The number of missing data for the socio-demographics should be split in half, because the

number of participants was 260. The number 520 represents replication of the participants

for the second market. Thus only 1 case of gender, 3 of education, 3 of household size, 1

working status, 5 frequency of buying groceries and 1 of buying PC/Laptops is actually

missing.

Appendix

376

Appendix 6

Figure 6.1 Scatterplots of the relationships with the behavioural variables.

The scatterplots indicate linear relationships.

Approach Avoidance Aminusa

Pleasure

Arousal

Appendix

377

Dominance

Similarity

Complexity

Source: this study

Appendix

378

Appendix 7. Research Questionnaire

(The format of the questionnaire is not identical to the online format- colors/ fonts and font

sizes are different).

Dear Sir/ Madam,

Thank you for considering taking this survey. This survey is part of a PhD research project

at Cardiff Business School. The aim is to examine your feelings and responses towards two

consumer shopping situations/markets. The completion of the survey should not take more

than 15 minutes of your time.

There are two parts in this survey. In the first part you will be asked some general, socio-

demographic information about yourself. In the second part, we will ask you to imagine you

are in two shopping situations and then ask some general questions about them. The two

situations are:

Situation 1: Grocery shopping and Situation 2: PC/Laptop shopping.

You do not need to be an expert buyer in any of the two situations in order to answer the

survey. Some levels of previous exposure are however required.

Situations are called 1 and 2 for convenience and practical reasons. You might be asked to

answer the questions of situation 2 first; this is not a problem. Please follow the instructions

and flow of the survey, which will guide you through the questions.

Appendix

379

Informed Consent (Cardiff Business School)

This survey is part of a PhD research project at Cardiff Business School. The aim is to

examine your feelings and responses towards two consumer shopping situations/markets.

The completion of the survey should not take more than 15 minutes of your time.

The anonymity and confidentiality of this survey is fully guaranteed and all data will be

stored securely. Your participation is completely voluntary and you can withdraw from the

research at any stage during the completion of the questionnaire. You do not need to

provide your name or address. You can, if you wish get a copy of findings of this research

by emailing me at [email protected] (after May 2013).

The data collected will be used for academic purposes only. If the study is published, the

information will be in aggregate form and it will not be identifiable as yours. Your data are

completely anonymous and are not possible to be deleted at a later point in time. You are

free to discuss any concerns with Dr. John Pallister (Cardiff Business School) or Ioanna

Anninou ([email protected]).

Your sincere responses are necessary to ensure the success of this research. Please try to

answer all questions.

By completing and submitting this survey I agree to take part in the study by Ioanna

Anninou (Cardiff Business School).

Appendix

380

Section A: About You.

The information about yourself will help us put your answers in context.

Your age (years)

18-24

25-34

35-44

45-54

55-64

65+

Your gender

Male

Female

Your higher (completed) education level

GCSE/ A-Levels

Vocational or Technical School

Higher Education (College, BSc)

Postgraduate Degree (MSc, MA, PhD)

Other. Please specify: ____________________

Your ethnic group

White

Mixed

Asian or Asian British

Black or Black British

Chinese or Chinese British

Other. Please specify: ____________________

Your household size

1 person

2 persons

More than 2 persons

Your working status

Employed full-time (30+ hours per week)

Employed part-time (less than 30 hours per week)

Self-employed

Housewife/husband

Unemployed

Student

Retired

I usually buy groceries

Online

In-store

I usually buy PC/Laptops

Online

In-store

Appendix

381

Section B: Your feelings and opinions of shopping situations.

Please take your time to think about the way you feel in each of the two situations and really

describe your feelings and opinions towards them; then rate your degree of agreement with

the questions that follow.

Remember that there are no correct or wrong answers. Please try to answer all questions.

Appendix

382

Situation 1: Grocery shopping.

For each pair of emotional words below, put a mark close to the word which you believe to

better describe your feelings when grocery shopping. Some of the pairs might seem

unusual, but you will probably feel more one way than the other.

In order to better describe your feelings you can imagine you are now in this situation.

Imagine that you are doing your weekly shopping at a large super market and you go

around the store with your trolley choosing any products you want from food to

household detergents. While you are doing this, YOU FEEL...

Happy 1 2 3 4 5 6 7 8 9 Unhappy

Annoyed 1 2 3 4 5 6 7 8 9 Pleased

Autonomous 1 2 3 4 5 6 7 8 9 Guided

Relaxed 1 2 3 4 5 6 7 8 9 Bored

Calm 1 2 3 4 5 6 7 8 9 Excited

Satisfied 1 2 3 4 5 6 7 8 9 Unsatisfied

Melancholic 1 2 3 4 5 6 7 8 9 Contented

Despairing 1 2 3 4 5 6 7 8 9 Hopeful

Frenzied 1 2 3 4 5 6 7 8 9 Sluggish

Awed 1 2 3 4 5 6 7 8 9 Important

Dull 1 2 3 4 5 6 7 8 9 Jittery

Aroused 1 2 3 4 5 6 7 8 9 Unaroused

Controlling 1 2 3 4 5 6 7 8 9 Controlled

Stimulated 1 2 3 4 5 6 7 8 9 Relaxed

Influenced 1 2 3 4 5 6 7 8 9 Influential

In-control 1 2 3 4 5 6 7 8 9 Cared-for

Sleepy 1 2 3 4 5 6 7 8 9 Wide-awake

Submissive 1 2 3 4 5 6 7 8 9 Dominant

Appendix

383


Please choose your response to each of the following questions regarding grocery shopping.

1. How much time would you like to spend on each grocery shopping trip?

None

A few minutes

Up to 30 minutes

Up to 60 minutes

More than one hour

Many hours

Many, many hours

2. How much would you try to get out of or avoid doing your shopping for groceries?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

3. Once in a grocery store, how much would you enjoy exploring around?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

4. Is grocery shopping a situation in which you might try to avoid other people, avoid

having to talk to them?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

5. To what extend is grocery shopping a situation in which you would feel friendly and

talkative to a stranger who happens to be near you?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

Appendix

384

6. How much would you try to avoid any looking around or exploring in a grocery store?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

Appendix

385


Please choose the appropriate circle to indicate your agreement with the following

statements.

Strongly

Agree Agree

Somewh

at Agree

Neither

Agree

nor

Disagree

Somewhat

Disagree

Disagr

ee

Strongly

Disagre

e

Some grocery

products are

so similar

that it is often

difficult to

spot new

products.

Some grocery

brands look

so similar

that it is

difficult to

detect

differences.

Most brands

in a grocery

store are very

similar and

are therefore

hard to

distinguish.

The more I

learn about

grocery

products, the

harder it gets

to choose the

best.

Appendix

386

Some grocery

brands look

so similar

that I don’t

know if they

are made by

the same

manufacturer.

There are so

many grocery

brands to

choose from

that I

sometimes

feel confused.

There are so

many stores

to shop from

that it is

sometimes

difficult to

decide where

to shop.

All the

information I

get on

different

grocery

products

confuses me.

Some grocery

products have

so many

features that a

comparison

of different

brands is

barely

possible.

Appendix

387

When buying

a grocery

product I

always feel

well

informed.

The

information I

get from

advertising

often is so

vague that it

is hard to

know what a

grocery

product

actually does.

When

purchasing

certain

grocery

products I

feel uncertain

about the

product

characteristic

s that are

particularly

important for

me.

To me there

are too many

products to

choose from

in a grocery

store.

Appendix

388



statements.

Strongly

Agree Agree

Somewhat

Agree

Neither

Agree

nor

Disagree

Somewhat

Disagree Disagree

Strongly

Disagree

Overall, I am

experienced in

shopping for

groceries.


How often do you shop for groceries?

Daily

2-3 times a week

Once a week

2-3 times a month

Once a month

Less than once a month

Never

Appendix

389

Situation 2: PC/Laptop shopping.

For each pair of emotional words below, put a mark close to the word which you believe to

describe better your feelings when shopping for computing products, like a PC/Laptop.

Some of the pairs might seem unusual, but you will probably feel more one way than the

other.

In order to better describe your feelings you can imagine you are now in this situation.

Imagine that you are in a computing store/website searching for a new PC or laptop.

While you are doing this, YOU FEEL...

Happy 1 2 3 4 5 6 7 8 9 Unhappy

Annoyed 1 2 3 4 5 6 7 8 9 Pleased

Autonomous 1 2 3 4 5 6 7 8 9 Guided

Relaxed 1 2 3 4 5 6 7 8 9 Bored

Calm 1 2 3 4 5 6 7 8 9 Excited

Satisfied 1 2 3 4 5 6 7 8 9 Unsatisfied

Melancholic 1 2 3 4 5 6 7 8 9 Contented

Despairing 1 2 3 4 5 6 7 8 9 Hopeful

Frenzied 1 2 3 4 5 6 7 8 9 Sluggish

Awed 1 2 3 4 5 6 7 8 9 Important

Dull 1 2 3 4 5 6 7 8 9 Jittery

Aroused 1 2 3 4 5 6 7 8 9 Unaroused

Controlling 1 2 3 4 5 6 7 8 9 Controlled

Stimulated 1 2 3 4 5 6 7 8 9 Relaxed

Influenced 1 2 3 4 5 6 7 8 9 Influential

In-control 1 2 3 4 5 6 7 8 9 Cared-for

Sleepy 1 2 3 4 5 6 7 8 9 Wide-awake

Submissive 1 2 3 4 5 6 7 8 9 Dominant

Appendix

390


Please choose your response to each of the following questions regarding your PC/laptop

shopping.

1. How much time would you like to spend on each PC/laptop shopping occasion?

None

A few minutes

Up to 30 minutes

Up to 60 minutes

More than one hour

Many hours

Many, many hours

2. How much would you try to get out of or avoid doing your shopping for PC/laptops?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

3. Once in a computing store/website, how much would you enjoy exploring around?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

Appendix

391

4. Is shopping for a PC/laptop a situation in which you might try to avoid other people,

avoid having to talk to them?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

5. To what extent is shopping for a pc/laptop a situation in which you would feel friendly

and talkative to a stranger who happens to be near you?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

6. How much would you try to avoid any looking around or exploring in a computing

store/website?

Not at all

Very slightly

Slight

Moderate

Much

Very much

Extremely so

Appendix

392



statements.

Strongly

Agree Agree

Somewhat

Agree

Neither

Agree

nor

Disagree

Somewhat

Disagree Disagree

Strongly

Disagree

Most

PC/laptops are

so similar that

it is often

difficult to

spot new

products.

Some

PC/laptop

brands look so

similar that it

is difficult to

detect

differences.

Most

PC/laptop

brands are

very similar

and are

therefore hard

to distinguish.

The more I

learn about

PC/laptops,

the harder it

gets to choose

the best.

Some

PC/laptop

Appendix

393

brands look so

similar that I

don’t know if

they are made

by the same

manufacturer.

There are so

many

PC/laptop

brands to

choose from

that I

sometimes feel

confused.

There are so

many

computing

stores/websites

to shop from

that it is

sometimes

difficult to

decide where

to shop.

All the

information I

get on

different

PC/laptops

confuses me.

PC/laptops

have so many

features that a

comparison of

different

brands is

barely

possible.

Appendix

394

When buying

a PC/laptop I

always feel

well informed.

The

information I

get from

advertising

often is so

vague that it is

hard to know

what a

PC/laptop can

actually

perform.

When

purchasing

PC/laptops I

feel uncertain

about the

product

characteristics

that are

particularly

important for

me.

To me there

are too many

PC/laptops to

choose from in

a computing

store/ website.

Appendix

395



statements.

Strongly

Agree Agree

Somewhat

Agree

Neither

Agree

nor

Disagree

Somewhat

Disagree Disagree

Strongly

Disagree

Overall, I am

experienced in

shopping for

PC/Laptops.


How often do you usually update/shop a PC/ Laptop or other computing equipment?

Once a month

Once every three months

Once every six months

Once a year

Once every two years

Less than once every two years

Never

We thank you for your time and cooperation in making this survey successful. The

main objective of this survey will be to examine consumers’ responses to retail

situations with a special interest on levels of consumer confusion.

If you agree to participate in this research,

please press the final next button to submit your survey.

Appendix

396

Appendix 8. Ethical Approval Form

Appendix

397

Appendix

398

Appendix

399